Are low-carbon cities and the development of the artificial intelligence industry contradictory?Evidence from China

Are low-carbon cities and the development of the artificial intelligence industry contradictory?Evidence from China | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Are low-carbon cities and the development of the artificial intelligence industry contradictory?Evidence from China luyuan tang, shiyao xie, yuan xu This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-6253945/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted 12 You are reading this latest preprint version Abstract The ‘low-carbon city’ pilot policy is a crucial measure for China to advance green and low-carbon development. The rapid growth of the artificial intelligence industry may lead to significant energy consumption and carbon emissions, prompting discussions about the potential conflict between low-carbon city construction and the growth of the artificial intelligence sector. This study utilizes China’s ‘low-carbon city’ pilot policy as a natural experiment, employs a staggered double difference model (Staggered DID), empirically tests the impact of the policy on the development of the urban artificial intelligence industry, and thoroughly explores its mechanisms and heterogeneity characteristics. The findings reveal that the low-carbon city pilot policy has not hindered the growth of the urban artificial intelligence industry. Compared to non-pilot cities, the policy's implementation has resulted in an average increase of approximately 29.4% in the size of the artificial intelligence industry in pilot cities. By examining the energy consumption characteristics across different segments of the artificial intelligence industry chain, this study identifies that the policy has a varied effect in promoting the development of the regional artificial intelligence industry by altering the urban energy consumption structure, primarily driving the clustering of the downstream segments of the low-emission industry chain. Additionally, enhancing the level of urban green innovation mainly supports the development of high-emission upstream segments and low-emission downstream segments of the industry. Heterogeneity analysis further indicates that the policy's positive effects significantly differ across various regions, city tiers, economic zones, and cities with differing levels of industrialization. This study not only confirms the compatibility of low-carbon city construction with the growth of the artificial intelligence industry but also offers vital policy insights for balancing economic growth with carbon emission targets. Social science/Economics Social science/Environmental studies Social science/Science technology and society Figures Figure 1 Figure 2 Introduction Climate change has become one of humanity's most significant challenges in the 21st century. In this global environmental governance effort, cities, as the primary centers of human activities, play a crucial role in achieving global climate objectives. Recent statistics show that urban areas contribute approximately 70% of global energy-related CO₂ emissions, and this proportion is on the rise (Luqman, Rayner, and Gurney, 2023 ). In China, urban carbon emissions account for 70–80% of total national emissions, with some studies estimating this figure could reach 85% by 2030 (Liguo Zhang et al., 2023 ). These statistics not only emphasize the vital role of cities in combating climate change but also highlight the significant challenges facing China, the world’s largest developing nation, in reaching its carbon neutrality targets. To advance the construction of ecological civilization and lead the way toward green, low-carbon development while achieving national greenhouse gas reduction targets, China’s National Development and Reform Commission introduced the Low-Carbon City Pilot Policy. Since launching the first batch of pilots in 2010, additional pilots were implemented in 2012 and 2017. This gradual expansion strategy for pilots reflects China’s unique approach to policy implementation. The selection of pilot cities comprehensively considered regional development stages, resource endowments, and differences in industrial structures, aiming to explore replicable regional low-carbon development models with diverse characteristics. In recent years, with the introduction of the ‘dual carbon’ goals and the execution of the Medium and Long-term Planning for Carbon Peak and Carbon Neutrality, low-carbon city construction has entered a new development phase, featuring an increasingly refined policy framework and strengthened implementation. The Low-Carbon City Pilot Policy, recognized as a significant innovation in environmental governance, has garnered extensive academic interest concerning its implementation effects and economic impacts. Existing research has highlighted the policy’s substantial influence across various dimensions: empirical studies have identified notable improvements in enterprise Total Factor Productivity (TFP) levels (X. Ren et al., 2022 ); in terms of innovation-driven development, the policy has fostered technological advancement through optimized resource allocation and reduced financing constraints (Zhu and Lee, 2022 ); regarding industrial structure, the policy has shown inhibitory effects on traditional high-carbon industries while simultaneously stimulating emerging sectors (C. Li et al., 2023 ). In enhancing energy efficiency, the policy mainly affects regional energy intensity through technological innovation rather than mechanisms for optimizing industrial structure (Hong, Chen, and Zhang, 2021 ). Additional research has examined the policy’s influence on carbon intensity and per capita carbon emissions, including potential impact mechanisms (H. Ren, Gu, and Zhou, 2022 ). However, these studies have largely centered on traditional industries or general economic indicators, with relatively limited exploration of emerging strategic industries, particularly artificial intelligence (AI) sector. The AI industry, as the cornerstone of the new generation of information technology revolution, plays a vital role in the development of China’s digital economy. In recent years, China has actively issued policy documents like the ‘New Generation Artificial Intelligence Development Plan’ and the ‘Three-Year Action Plan for Promoting New Generation Artificial Intelligence Industry Development,’ elevating AI industry growth to a matter of national strategy. However, the AI industry demonstrates unique dual characteristics: on one hand, AI technology can optimize energy systems and enhance resource utilization efficiency, contributing to environmental governance; on the other hand, the extensive computing facilities and data center operations involved in industry growth generate substantial carbon footprints. Research shows that global energy-related carbon dioxide emissions rose by 1.1% in 2023, reaching a historic high of 37.4 billion tons (International Energy Agency, 2023 ). This duality complicates the relationship between low-carbon city policies and AI industry development, revealing its complexity and nuance. Current academic research on the relationship between artificial intelligence and carbon emissions is still in its early stages. While some scholars have demonstrated that AI applications can effectively promote carbon reduction (Tao et al., 2024 ), the influence of low-carbon city policies on the development of the AI industry has not been systematically examined. This research gap makes it challenging to ascertain whether environmental policies hinder the growth of emerging industries, how to reconcile emission reduction targets with the needs of industrial development, and whether policy design should take into account industry-specific characteristics. This study investigates the impact of low-carbon city policies on the AI industry, with significant theoretical and practical implications. On the theoretical level, the research extends the industrial impact theory of environmental regulation, enhances understanding of constraints on emerging industry development, and perfects the theoretical framework for coordinating environmental and industrial policies. On the policy level, the research findings provide empirical evidence for optimizing low-carbon city policy design, aiding in balancing environmental protection with industrial development, and offering references for formulating differentiated industry support policies. On the practical level, the findings will assist city administrators in better understanding policy impacts, guiding enterprises in their green transformation, and promoting the sustainable development of the AI industry. The innovation of this study lies in its pioneering systematic examination of the impact of environmental policies on strategic emerging industries, filling a gap in related research fields, and providing crucial guidance for facilitating China’s green, low-carbon transition and the development of its digital economy. The subsequent chapters are organized as follows: Section 2 introduces the policy background and presents the research hypotheses; Section 3 describes the research design; Section 4 reports the main results; Section 5 presents the mechanism analysis results; Section 6 introduces the heterogeneity analysis; and finally, Section 7 draws the research conclusions based on the previous analyses and provides theoretical implications and policy recommendations. Research background and hypotheses development Policy background. The implementation of low-carbon policies represents China’s strategic choice in addressing rapidly increasing carbon emissions and promoting green transformation. In 2005, China became the world’s largest carbon emitter, with emissions reaching 6.098 billion tons, surpassing the United States’ 5.880 billion tons. Faced with growing international pressure for emission reductions and domestic environmental constraints, China urgently needed to transform its development model. At the 2009 Copenhagen Climate Conference, China pledged to reduce its carbon intensity by 40–45% by 2020 compared to 2005 levels, marking China’s transition from an extensive to a low-carbon development model. In this context, China launched three rounds of low-carbon city pilot programs in July 2010, November 2012, and February 2017. The first round consisted of five provinces (Guangdong, Liaoning, Hubei, Shaanxi, and Yunnan) and eight cities (Tianjin, Chongqing, Shenzhen, Xiamen, Hangzhou, Nanchang, Guiyang, and Baoding). The second round included 28 cities, such as Beijing, Shanghai, and Qinhuangdao, along with Hainan Province. The third round, building on the successes of the previous two, covered 45 cities (districts and counties). Due to the extensive scope of the third round, pilot regions were required to develop replicable and scalable experiences, with successful practices gradually being promoted nationwide by 2020. Therefore, this study considers 2010, 2012, and 2017 as the implementation time points for the low-carbon city pilot policy. Hypotheses development: The impact of Low-carbon city on artificial intelligence industry development. Existing literature presents two competing explanations for the relationship between environmental regulation and industrial upgrading: the traditional school argues that environmental regulation fosters the upgrading of industrial structure through the factor relocation effect (Zhaohui Chong et al., 2017 ), while the revisionist school highlights that the institutional environment may create an upgrading barrier effect (Antoniou, F., & Mageiropoulos, T., 2024 ). This theoretical divergence offers a distinct perspective on understanding the AI industry effect of the low-carbon city pilot policy. From the theoretical perspective of the revisionist school, within a specific institutional context, environmental regulation may create a ‘compliance burden.’ To fulfill short-term environmental compliance requirements, enterprises allocate limited resources to end-of-pipe governance instead of pursuing fundamental technological innovation (Wu, G et al., 2024 ), which leads to insufficient motivation for innovation. Simultaneously, under the cost pressures of environmental regulation, enterprises may prioritize established technology paths over cutting-edge exploration, thereby resulting in a ‘low-end lock-in’ effect (Wang, X et al., 2022 ). This structural constraint presents a potential hindrance to the development of emerging industries. However, China’s institutional context may undermine the theoretical effects outlined above. As a large developing country, China often takes a gradual reform approach to environmental regulation, using a ‘pilot first’ strategy that grants local governments greater autonomy in policy implementation (Chen, H et al., 2024 ). In this framework, environmental regulation generally pairs with supporting industrial policies, creating a dual policy combination of ‘environmental constraints + innovation incentives’. As a comprehensive reform initiative, the low-carbon city pilot not only establishes carbon emission constraints but also offers innovation incentives such as fiscal subsidies, tax breaks, and technical support, which effectively mitigates the potential innovation-inhibiting effects that might arise from basic environmental regulations. Furthermore, under the performance evaluation system, local governments in China often leverage environmental regulation as a catalyst for industrial upgrading, facilitating the concentration of high-tech industries in their regions through a mix of investment promotion and industrial initiatives policies. Based on the theoretical analysis and considerations of real-world situations outlined above, this study proposes the following core hypothesis: H1: The low-carbon city pilot policy can foster the development of the artificial intelligence industry. The effect of the low-carbon city pilot policy on the artificial intelligence industry primarily operates through the cost transmission mechanism, shaped by environmental regulation constraints, and the knowledge spillover mechanism associated with technological externalities. According to the new economic geography theory (Krugman, 1992 ) and technology spillover theory, environmental regulation can enhance industrial development advantages through the dual mechanisms of optimizing energy structures and promoting green technology innovation. Hypotheses development: The mediating role of energy consumption structure. As the central subject of carbon emissions within the urban economic system, the low-carbon transformation of enterprises directly influences the effectiveness of urban environmental governance. From a cost-based theoretical perspective, environmental regulatory policies have raised the prices of traditional energy sources, encouraging artificial intelligence companies to seek lower-cost energy alternatives. This shift optimizes their green energy consumption structure and significantly enhances the geographical clustering of the industry. Specifically, on one hand, the low-carbon city pilot policy has notably increased the costs associated with traditional energy sources like coal-fired power through mechanisms such as carbon emission rights trading, environmental taxes, and differentiated electricity pricing. On the other hand, the policy has spurred the growth of renewable energy capacity and technological advancements through the green power quota system and priority dispatch mechanism, leading to a continuous reduction in the marginal costs of clean energy sources like wind and solar power. As the supply of green energy reaches a critical scale, its low marginal cost characteristics will reshape the regional power market dynamics, resulting in a competitive decline in low-carbon energy prices. This structural change exerts a significant gravitational influence on the artificial intelligence industry, which has high electricity demands. When pilot areas develop an electricity pricing advantage, both upstream and downstream enterprises within the artificial intelligence industry chain will naturally form spatial clusters based on cost efficiency and operational stability. Accordingly, this paper proposes: H2: The low-carbon city pilot policy attracts the agglomeration and development of the artificial intelligence industry by changing the energy consumption structure. Hypotheses development : The mediating role of green technology innovation. In addition to influencing industrial agglomeration by altering the energy consumption structure, environmental regulation can also impact the path of technology diffusion by stimulating innovation activities. From the perspective of innovation compensation, moderate environmental regulation can encourage enterprises to pursue green technology innovation, and its knowledge spillover effect can lower the technology adoption threshold for related industries, thereby facilitating the industry's growth. Specifically, there are notable differences between green technology innovation and traditional technology innovation, as green innovation often exhibits stronger universality and positive externality characteristics (Hu Y et al., 2022 ). For instance, technologies for improving energy efficiency and carbon capture and storage have application principles and solutions that can be transferred across industries, making green innovation more likely to produce knowledge spillover effects. Conversely, the artificial intelligence industry is significantly technology-intensive, requiring continuous integration and application of cross-domain technologies. Green technology innovation triggered by environmental regulation, particularly breakthroughs in energy efficiency optimization, smart grids, and distributed energy management, aligns well with the technological needs of the artificial intelligence industry. This technological connection allows the knowledge spillover of green innovation to reduce the technology adoption threshold for artificial intelligence companies. From the perspective of technological economics, reducing technology adoption costs directly impacts the industry's development trajectory. The knowledge spillover effect of green technology innovation not only lowers the R&D costs for artificial intelligence companies but also provides them with new application scenarios and business models, further promoting industrial development. Based on this, this paper proposes: H3: The low-carbon city pilot policy triggers green technology innovation and enables the growth of the artificial intelligence industry. Research design The definitions and the descriptive statistics of the variables utilized in this study are given in Table 1 . Table 1 Descriptive statistics VARIABLES N Mean Sd Min Max DAII 4,543 4.576 1.921 0.693 9.707 LCPP 4,543 0.300 0.458 0 1 Size 4,543 5.872 0.672 3.822 7.116 Dens 4,543 0.271 0.461 0.00527 3.026 Gdp 4,543 10.58 0.675 8.919 12.05 Fina 4,543 2.383 1.140 0.896 6.760 Open 4,543 0.180 0.283 0.00165 1.657 Edu 4,543 0.0185 0.0236 0.000629 0.117 Inf 4,543 0.815 0.374 0.171 2.245 Dependent Variable: Development of artificial intelligence industry (DAII). To measure the development levels of the AI industry, following the methodology of the Hong Kong Productivity Council (2024) and Sun et al. ( 2022 ), we utilized Python web crawling technology to perform fuzzy keyword matching queries on the ‘business scope’ and ‘company name’ fields in the Qichacha database. We employed keywords related to AI applications, including ‘artificial intelligence,’ ‘cloud,’ ‘data,’ ‘IoT,’ and ‘machine learning.’ The data was aggregated by year and region to create panel data of AI enterprises in Chinese cities from 2007 to 2022. The natural logarithm of the number of AI enterprises in each city-year observation was employed to measure AI industry development (DAII). Independent Variable: Low-carbon city pilot policy (LCPP). We employed the staggered Difference-in-Differences (DID) method to examine the implementation effects of the low-carbon city pilot policy (Popp, 2006 ; Pei et al., 2019 ). When the pilot region was a province, all cities within that province were regarded as pilot cities. If the same city appeared in different batches of pilot lists, the earliest batch was utilized to establish the pilot start time. Accordingly, we constructed a policy dummy variable (LCPP) as follows: $$\:LCP{P}_{i,t}=Trea{t}_{i}\times\:Tim{e}_{t}\:\:\:\:\:\:\:\:\:$$ 1 Where \(\:Trea{t}_{i}\) and \(\:Tim{e}_{t}\) are dummy variables, \(\:i\) represents city \(\:i\) and \(\:t\) represents year \(\:t\) . \(\:Trea{t}_{i}\) equals 1 when the city is the pilot city, and 0 otherwise. \(\:\text{T}\text{i}\text{m}{\text{e}}_{\text{t}}\) takes the value of 1 when the policy is actually, and 0 otherwise. Control variables. Following previous research (S. Li et al., 2024 ), we controlled for several city characteristics that might influence the development of the AI industry. These variables include: (1) the natural logarithm of year-end total population, representing population size (Size); (2) GDP divided by administrative land area, representing urban economic density (Dens); (3) the natural logarithm of GDP per capita, representing economic development level (Gdp); (4) year-end financial institution loan balance divided by GDP, representing the degree of financial development (Fina); (5) total import-export volume divided by GDP, representing the degree of openness (Open); (6) the number of students in higher education institutions per year-end total population, representing education level (Edu); and (7) fixed asset investment divided by GDP, representing infrastructure level (Inf). Model settings. The DID method is the most widely recognized model in policy evaluation literature, as it significantly mitigates omitted variable bias and accurately identifies exogenous policy shocks. Considering the three batches of low-carbon city pilots, we adopted a ‘staggered DID’ approach to study the impact of low-carbon city construction on AI industry development. $$\:DAI{I}_{i,t}=\alpha\:+\beta\:LCP{P}_{i,t}+\lambda\:\sum\:Control{s}_{i,t}+{\gamma\:}_{t}+{\mu\:}_{i}+{\epsilon\:}_{i,t}$$ 2 Where \(\:i\) and \(\:t\) denote city and year respectively. \(\:DAI{I}_{i,t}\) denotes development of artificial intelligence industry. \(\:LCP{P}_{i,t}\) is a policy dummy variable, and \(\:\beta\:\) shows how the LCPP affects DAII. \(\:Controls\:\) denotes a series of city characteristics controlled for in this paper. \(\:{\gamma\:}_{t}\) and \(\:{\mu\:}_{i}\) denote the fixed effect of year and city respectively, and \(\:{\epsilon\:}_{i,t}\) is a random interference item. Data sources. The study covers data from 2007 to 2022, including 285 prefecture-level and higher cities in China (excluding those with significant data gaps). Data on AI industry development was obtained from the Chinese enterprise credit information platform, ‘Qichacha.’ The low-carbon city pilot list was sourced from relevant documents of the National Development and Reform Commission. Additional data came from the Express Professional Superior (EPS) platform and the ‘China City Statistical Yearbook.’ Linear interpolation was employed to fill in minor missing data in some cities. Results and discussion Main results. Based on the aforementioned benchmark model, this section examines the quantitative impact of implementing the low-carbon city pilot policy on the development of the urban artificial intelligence industry. The estimated results are presented in Table 2 . The coefficient of the LCPP is the focus of our analysis. In column (1) of Table 2 , without the addition of control variables and fixed effects, the LCPP coefficient indicates a positive effect at the 1% significance level, which preliminarily verifies the policy's effectiveness. Subsequently, after introducing both the city-year two-way fixed effects and all control variables in columns (2) and (3), the coefficient of the core explanatory variable still holds 1% statistical significance, demonstrating the robustness of the main results. For each 1 percentage point increase in the LCPP policy, the DAII increases by 1.39 percentage points. This quantitative finding effectively confirms research hypothesis H1, which states that the low-carbon city pilot policy can promote the development of the urban artificial intelligence industry. Table 2 Main results VARIABLES (1) (2) (3) DAII DAII DAII LCPP 1.391*** 0.699*** 0.156*** (23.75) (13.91) (6.59) Size 1.038*** (65.09) Dens 0.250*** (8.59) Gdp 1.323*** (52.10) Fina 0.297*** (21.17) Open 0.465*** (8.96) Edu 7.144*** (10.91) Inf 0.165*** (5.38) Constant 4.159*** 3.212*** -17.565*** (129.62) (27.64) (-61.76) Observations 4,543 4,543 4,543 R-squared 0.110 0.487 0.885 id fe no yes year fe no yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance levels at 1%, 5%. and. 10%, respectively. Parallel trend test. The prerequisite for the DID method is that both the experimental and control groups meet the parallel trends assumption prior to the policy implementation. This means there are no systematic differences in AI industry development between pilot and non-pilot cities before the introduction of the low-carbon city pilot policy. Eq. ( 3 ) is structured to examine the dynamic trends of digital government policy (Beck et al., 2010 ; Bertrand and Mullainathan, 2003 ), as follows: $$\:DAI{I}_{i,t}={\beta\:}_{0}+{\sum\:}_{k=-5}^{10}{\beta\:}_{k}LCP{P}_{i,t}^{k}+{\beta\:}_{2}Control{s}_{i,t}++{\gamma\:}_{t}+{\mu\:}_{i}+{\epsilon\:}_{i,t}$$ 3 To provide a more intuitive view of the results from the parallel trend test, we plotted the estimated values of the coefficients along with their 95% confidence intervals. The results of the parallel trend test are illustrated in Fig. 1 below. Prior to the policy's implementation, the regression coefficient for LCCP did not demonstrate statistical significance, indicating that there was no notable difference in carbon emission efficiency between pilot cities and non-pilot cities before the policy intervention. Following the policy implementation, the LCCP coefficient began to show a continuous upward trend. This trend aligns with the main regression results, suggesting that the policy may positively influence carbon emission efficiency. The increase in the coefficient suggests that the improvement in carbon emission efficiency is notably greater in the intervened cities compared to the non-intervened cities. Notably, one year after the policy was enacted, the coefficient achieved statistical significance, indicating a potential one-year lag effect of LCCP. This lag effect may arise from the delay between policy planning and execution, or it could reflect the time cities need to adjust and respond to policy requirements. Placebo test. Although this study has eliminated as many confounding factors as possible that might influence the results, such as including city and time fixed effects in the regression, these measures alone are insufficient to overcome the problems arising from omitted variable bias. Therefore, we further conduct robustness checks through placebo tests. Following Chetty et al. ( 2009 ), we randomly select a group of cities as pseudo-pilot cities, reconstruct the policy dummy variable, and obtain estimation coefficients. The results of 500 counterfactual estimations are shown in Fig. 2 . The pseudo-regression coefficients are concentrated around zero, approximating a normal distribution, while the baseline regression estimation coefficient falls outside this distribution. Therefore, we can claim that our baseline estimation results in this article are not swayed by unobservable factors, stipulating that the findings of this article are replicable. Endogeneity analysis. Reverse causality is one of the significant reasons for the endogeneity problem. Specifically, the low-carbon city pilot policy is not entirely exogenous, and the developmental needs of the regional artificial intelligence industry may in turn promote the implementation of the policy, creating a two-way causal relationship. Therefore, this paper employs the instrumental variable method to address the potential endogeneity issue in the benchmark regression. This paper constructs two sets of geographical-historical instrumental variables. First, this paper follows the approach of Zhou (2024) to manually collect government work reports from 285 prefecture-level cities from 2008 to 2022, segment the text of the reports, count the frequency of words related to environmental regulation and environmental protection, and calculate the proportion of these words in relation to the total word frequency of the full text of the government report as an instrumental variable (IV1). On one hand, the frequency of words associated with environmental regulation and protection in the government work report indicates the level of attention local governments give to environmental issues. This level of attention is to some extent independent of the implementation effect of the low-carbon city pilot policy. Therefore, this instrumental variable can be regarded as meeting the exclusivity requirement. On the other hand, this instrumental variable is highly correlated with the implementation of the low-carbon city pilot policy, as greater attention from local governments to environmental matters increases the likelihood of actively promoting the low-carbon city pilot policy, thus fulfilling the correlation requirement and qualifying as a suitable instrumental variable. The low-carbon city pilot policy stimulates green technology innovation and facilitates the development of the artificial intelligence industry. Second, we refer to the research of Dong Zhiqinghua and Wang Hui (2019) and construct the industrial pollution stock index before the policy implementation as an instrumental variable (IV2): ① Select six major pollutants such as SO₂, smoke, and COD; ② Use the entropy method to determine the weights: \(\:{\text{W}}_{\text{j}\text{t}}=\frac{1-{\text{E}}_{\text{j}\text{t}}}{{\sum\:}_{\text{k}=1}^{6}(1-{\text{E}}_{\text{k}\text{t}})}\) , where \(\:{\text{E}}_{\text{j}\text{t}}\) = \(\:-{\sum\:}_{\text{i}=1}^{\text{N}}\frac{{\text{E}}_{\text{i}\text{j}\text{t}}}{{\text{E}}_{\text{j}\text{t}}}\text{l}\text{n}\frac{{\text{E}}_{\text{i}\text{j}\text{t}}}{{\text{E}}_{\text{j}\text{t}}}\) ;③ Calculate the standardized pollution intensity in batches: \(\:{\text{P}\text{I}}_{\text{i}\text{t}}^{\text{b}}\) = \(\:\sum\:_{\text{j}=1}^{6}(\frac{{\text{E}}_{\text{i}\text{j}\text{t}}/{\text{G}\text{D}\text{P}}_{\text{i}\text{t}}}{\frac{1}{{\text{N}}_{\text{b}}}{\sum\:}_{\text{i}=1}^{{\text{N}}_{\text{b}}}{\text{E}}_{\text{i}\text{j}\text{t}}/{\text{G}\text{D}\text{P}}_{\text{i}\text{t}}}\) ) \(\:\times\:{\text{W}}_{\text{j}\text{t}}\) (representing the number of candidate cities in batch b). On one hand, from the correlation perspective, the stock of industrial pollution prior to the policy's implementation can reflect the severity of environmental issues in a region. This severity often influences local government’s attention to and execution of the low-carbon city pilot policy. Therefore, this instrumental variable is strongly correlated with the low-carbon city pilot policy. On the other hand, from the exogeneity perspective, the industrial pollution stock before the policy's implementation is established prior to the policy itself and is unrelated to the economic effects that occur post-implementation. Therefore, this instrumental variable satisfies the exogeneity requirement, indicating its effectiveness. This paper employs the two instrumental variables mentioned above for two-stage estimation, with specific regression results presented in Table 3 . We report the estimation outcomes of IV1 and IV2 in columns (1) through (4). The results indicate that in the first stage, the first instrumental variable significantly positively impacts LCPP at the 5% level, while the second instrumental variable shows a significant positive impact on LCPP at the 1% level, confirming our analysis. In the second stage, both instrumental variables are introduced individually, and the estimated coefficients of LCPP on DAII are significantly positive at the 1% level. The model passes both the unidentification test and weak identification test, suggesting that the two instrumental variables developed in this paper are sufficiently exogenous and that there is no weak instrumental variable issue. As shown in Table 3 , the estimation results of this paper remain unaffected by the endogeneity problem, maintaining robustness. Table 3 Endogeniety analysis VARIABLES (1) (2) (3) (4) Phase 1 Phase 2 Phase 1 Phase 2 IV1 IV1 IV2 IV2 LCPP 3.570*** 4.009*** (4.68) (4.50) IV1 0.151** (2.08) IV2 0.164*** (4.82) Controls yes yes yes yes Observations 4,543 4,543 4,543 4,543 id fe yes yes yes yes year fe yes yes yes yes R-squared 0.107 0.112 0.111 0.114 Kleibergen-Paap rk LM 70.4547*** 69.5295*** Kleibergen-Paap rk Wald F 69.110 71.820 Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance levels at 1%, 5%. and. 10%, respectively. Propensity score matching (PSM)-DID. To control for potential sample selection bias, we employ the PSM-DID method for re-estimation (Fei & Zhang, 2021). This study uses nearest neighbor matching for verification. Using control variables as matching variables shown in Table 4 , before matching, the ATT (Average Treatment Effect on the Treated) is 5.5300 and significant at the 1% level (absolute t-value greater than 2.58); after matching, the ATT is 0.8542, also significant at the 1% level. In the balance test shown in Table 4 , the %bias of matching variables is less than 10% and notably smaller than the pre-matching %bias, indicating no significant differences between the control and experimental groups after PSM. Table 4 Balancing test results of PSM Variables Unmatched Matched Mean %Reduction t-test V(T)/V(C) Treated Control %Bias [Bias] t p > t Size U 5.889 5.866 3.3 72.6 1.00 0.316 0.94 M 5.883 5.877 0.9 0.22 0.825 0.77* Dens U 0.542 0.196 37.7 90.0 14.57 0.000 19.09* M 0.448 0.413 3.8 1.31 0.191 1.56* Gdp U 10.875 10.448 66.3 99.5 19.95 0.000 0.75* M 10.863 10.865 -0.3 -0.09 0.925 0.92 Fina U 2.805 2.223 45.6 99.4 14.87 0.000 1.67* M 2.783 2.789 -0.3 -0.05 0.959 0.43* Open U 0.242 0.161 24.8 99.7 8.04 0.000 1.60* M 0.232 0.231 0.1 0.02 0.986 0.94 Edu U 0.025 0.016 37.5 65.0 12.47 0.000 2.05* M 0.025 0.022 13.1 3.20 0.001 1.48* Inf U 0.813 0.808 1.3 61.3 0.39 0.696 1.20* M 0.813 0.811 0.5 0.12 0.901 1.14* Notes: The standard errors are indicated in parentheses. ⁎⁎⁎, ⁎⁎, and ⁎ denote significance levels at 1%, 5% and 10%, respectively. Table 5 presents the PSM-DID estimation results for DAII, with significantly positive estimation coefficients, validating the robustness of our main results. Table 5 PSM-DID estimation VARIABLES DAII VARIABLES DAII LCPP 0.083*** Edu -8.124*** (5.43) (-4.19) Size 2.450*** Inf 1.006*** (4.42) (5.21) Dens 5.096*** Constant -10.523*** (16.54) (-2.62) Gdp -0.787*** Observations 4,420 (-5.15) R-squared 0.987 Fina 0.044 id fe yes (0.76) year fe yes Open 0.616** (2.00) Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and 10%, respectively. Substitution of explained variable. From the sample observation, it is found that the business scope provided at the time of enterprise registration is relatively broad, leading to excessive granularity when directly capturing and identifying artificial intelligence enterprises based on their name or business scope. Therefore, to more accurately define and analyze the artificial intelligence industry, we utilize the national standard classification code to categorize the ten industries of software development, technology promotion services, other science and technology promotion services, other information technology services, engineering and technology research and experimental development, information technology consulting services, consulting and investigation, Internet information services, information system integration, and Internet of Things technology services as core artificial intelligence enterprises, while classifying the remaining industries as marginal enterprises. This study employs the number of regional core artificial intelligence enterprises to represent the scale of the regional artificial intelligence industry as an alternative dependent variable. The regression results from Table 5 (1) indicate that the impact coefficient is 0.262, which is significant at the 1% level, supporting the hypothesis H1. Table 6 Robustness tests: other methods VARIABLES (1) (2) (3) (4) LCPP 0.262*** 0.495*** 0.551*** 0.316*** (3.42) (2.68) (5.71) (5.73) Observations 4,543 4,543 4,259 4,259 R-squared 0.260 0.497 0.718 0.509 id fe yes yes yes yes year fe yes yes yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and 10%, respectively. Changing clustering method. To account for clustering of standard errors, we clustered them at the province and year levels for robustness checks. As shown in Table 6 Column (2), the estimated coefficient for low-carbon city construction remains significant at the 1% level, confirming the robustness of the baseline regression results. Considering implementing lag one-stage. Given that the impact of the LCPP intervention might not be immediate, we lagged the core regressor (LCCP) by one period. We applied the same method to all control variables to address simultaneous equation bias. As shown in Table 6 , Column (3), the LCPP policy demonstrates a positive promotional effect on the development of the urban AI industry at the 1% significance level. This finding is consistent with the baseline regression results, further reinforcing our previous conclusions. Eliminating the influence of COVID-19. To ensure the accuracy of these findings, it was essential to eliminate the impact of COVID-19 in 2020, as it could significantly affect data distribution, trends, and outliers. Neglecting to account for this influence could lead to biased or inaccurate results. Therefore, this study excluded data from 2020 to remove the effects of COVID-19. Column (4) of Table 6 shows that after controlling for COVID-19 effects, the LCCP policy estimation remains significant at the 1% level, demonstrating the robustness of the baseline regression model. Mechanism analysis:Energy consumption structure. According to the theoretical analysis presented in the previous article, the low-carbon city pilot policy may facilitate the agglomeration development of the artificial intelligence industry by optimizing the energy consumption structure. To verify this mechanism, the proportion of regional coal consumption—directly related to the energy consumption structure—is used here to represent the regional energy consumption structure for testing hypothesis 2. Column (1) of Table 7 displays the estimated results of the impact of the low-carbon city pilot policy on the energy consumption structure. The results reveal that the coefficient estimate of the LCPP is -0.112, and it passes the significance test at the 1% level. This indicates that, under cost theory, the low-carbon city pilot policy significantly reduces the intensity of regional use of traditional energy. This paper further examines the impact of energy consumption structure on the development of the artificial intelligence industry. The results in column (2) indicate that the coefficient estimate of energy consumption structure is -0.037, which also passes the significance test at the 1% level. This suggests that the low-carbon city pilot policy can promote the development of the regional artificial intelligence industry by decreasing the intensity of regional consumption of traditional energy, such as coal. Table 7 Mechanism analysis: energy consumption structure (1) (2) (3) (4) (5) VARIABLES NY DAII Upstream Midstream Downstream LCCP -0.112*** (-3.07) NY -0.037*** -0.082 0.032* -0.024*** (-2.64) (-1.14) (1.62) (-1.68) Controls yes yes yes yes yes Observations 4,543 4,543 4,543 4,543 4,543 R-squared 0.883 0.367 0.349 0.139 0.171 id fe yes yes yes yes yes year fe yes yes yes yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and 10%, respectively. This paper also examines the impact of the energy consumption structure on various aspects of the artificial intelligence industry chain. Referring to the industry-level division paradigm officially proposed by Qichacha, the artificial intelligence industry chain is divided into three segments: the basic layer (upstream link), the technical layer (midstream link), and the application layer (downstream link). Specifically, the basic layer includes the three pillars of computing power infrastructure, core hardware components, and the data resource system; the technical layer focuses on key technical modules such as algorithm research and development, platform construction, and tool development; the application layer concerns the development of smart terminal products and the implementation of industry scenario solutions. In identifying the categories of sample enterprises, this paper constructs a text classification algorithm model based on the QWEN-7B large language model (Bai et al., 2023 ) and leverages natural language processing technology to extract semantic features and recognize patterns in the business scope text of the enterprise, ultimately achieving precise positioning and hierarchical attribution determination of the sample enterprises within the industrial chain. By analyzing data from sample enterprises, it is observed that the upstream sector primarily comprises high-emission companies dependent on traditional energy sources. Since basic hardware manufacturing and large-scale data center operations necessitate a continuous and stable supply of significant amounts of electricity, they predominantly rely on traditional energy sources, such as coal-fired power. The midstream sector can partially utilize renewable energy for technological research and development, as well as platform operations, but it still requires assurances from traditional energy sources for computing-intensive tasks. The downstream sector mainly consists of low-emission enterprises that utilize clean energy, since terminal applications have relatively low energy demands, and the companies' images and ESG requirements compel them to use clean energy to meet carbon neutrality objectives. Columns (3), (4), and (5) of Table 7 display the estimated effects of the regional energy consumption structure on the upstream, midstream, and downstream segments of the industry. The findings reveal that the energy consumption structure significantly negatively impacts the downstream sector at the 1% level, while it positively affects the midstream sector at the 10% level. This indicates that the low-carbon city pilot policy fosters the growth of the regional artificial intelligence industry by optimizing the energy consumption structure. It mainly supports the development of the downstream portion of the industrial chain while hindering the growth of the midstream part. This may be due to significant technological substitution effects and the rigidities of factor substitution in the energy transformation process: downstream application scenarios can directly align with terminal market demands and quickly achieve clean energy substitution through flexible energy adaptations (such as deploying distributed computing power), thereby benefiting from policy-driven technological diffusion gains. In contrast, midstream technology-focused companies are restricted by the path dependency of high-precision algorithm research and development on a stable energy supply, facing dual pressures from equipment upgrades and rising R&D costs in the short term, which leads to the variability of policy transmission. Mechanism Analysis: Green Technology Innovation. To test the potential mechanism behind the innovation compensation effect, this paper evaluates whether the low-carbon city pilot policy can stimulate the growth of the regional artificial intelligence industry by promoting green technology innovation. Green technology innovation may manifest as the low-carbon transformation of production processes, the upgrading of clean technology equipment, and the reconstruction of green product systems. The low-carbon city pilot policy encourages enterprises to move beyond traditional technology pathways and adopt green innovation by establishing a carbon quota trading mechanism and a tiered carbon tax system. This includes developing intelligent energy consumption management systems, deploying industrial Internet of Things (IIoT) energy efficiency optimization platforms, and implementing carbon footprint real-time monitoring technology. Considering the availability of urban green innovation technology data, this paper uses the logarithm of the number of green inventions filed by the city in that year as a proxy variable for green innovation technology. This indicator not only reflects the knowledge output density of technological innovation but also addresses the quality heterogeneity problem associated with utility model patents (Aghion et al., 2016 ). Table 8 , column (1) presents the estimated results concerning the impact of the low-carbon city pilot policy on green technology innovation. The coefficient estimate for LCPP is 0.161, and it passes the significance test at the 5% level. This indicates that the low-carbon city pilot policy has significantly stimulated green technology innovation. Furthermore, this paper extends the analysis to investigate the impact of green technology innovation on the development of the regional artificial intelligence industry. Table 8 , columns (2), (3), (4), and (5) illustrate that the promotional effects of green technology innovation on the entire industrial chain, the upstream basic layer, and the downstream application layer are 0.035 (p < 0.1), 0.021 (p < 0.1), and 0.047 (p 0.1). This may be attributed to the ‘double-end driving effect’ of technological innovation on the industrial chain: upstream basic layer enterprises have significantly reduced their marginal emission reduction costs (MAC) due to policy subsidies and tax incentives (such as subsidies for high-efficiency chip design), which directly support clean technology research and development. In contrast, downstream application layer enterprises can quickly integrate green technology into smart terminal products (such as new energy autonomous driving systems) through flexible resource reconfiguration, allowing them to gain market premiums via product differentiation. Conversely, the midstream technology layer faces limitations due to the path-dependence characteristics of algorithm models (Arthur, 1989 ), making it difficult to break out of the existing technology framework and achieve green transformation in the short term, which hinders the transmission of innovation dividends. Table 8 Mechanism analysis: Green Technology Innovation (1) (2) (3) (4) (5) VARIABLES GTI DAII Upstream Midstream Downstream LCCP 0.161** (2.49) GTI 0.035*** 0.057*** 0.004 0.021* (6.80) (3.17) (0.80) (1.67) Controls yes yes yes yes yes Observations 4,543 4,543 4,543 4,543 4,543 R-squared 0.260 0.884 0.601 0.114 0.136 id fe yes yes yes yes yes year fe yes yes yes yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and 10%, respectively. Heterogeneity analysis: Urban development level. Cities at various stages of development showcase differences in socioeconomic factors like population size, economic base, industrial structure, and capacity for policy implementation. Therefore, this study explores the diverse impacts of low-carbon city pilot policies across two dimensions: city tiers and location within the Yangtze River Economic Belt. Based on the ‘2022 City Commercial Attractiveness Rankings’ by Yicai, we categorize the sample cities into three groups - first-tier, second-tier, and third-tier or below - for heterogeneity analysis. Additionally, given the strategic significance of the Yangtze River Economic Belt, we further divide the sample into groups for the Yangtze region and non-Yangtze region for supplementary testing. Columns (1)-(3) of Table 9 present regression results for first-tier, second-tier, and third-tier cities, respectively. The findings indicate that low-carbon city pilot policies significantly promote the development of the AI industry in third-tier and below cities, whereas the effects on first- and second-tier cities are insignificant. Columns (4) and (5) of Table 9 display results for cities within and outside the Yangtze Economic Belt, revealing significant positive effects for cities within the Belt but insignificant effects for those beyond it. This may be due to the fact that, compared to the established industrial systems in first- and second-tier cities, as well as in Yangtze Economic Belt cities, emerging industries like AI in lower-tier and non-Belt cities are relatively smaller. This smaller scale allows policy incentives to concentrate limited resources for rapid industrial clustering, thereby facilitating a quick breakthrough of scale economy thresholds and enabling rapid industrial development. Table 9 Heterogeneity analysis: Urban development level VARIABLES (1) (2) (3) (4) (5) LCPP 0.202 0.403 0.034** 0.296 0.526* (0.24) (1.12) (2.27) (1.47) (1.86) Controls yes yes yes yes yes Observations 374 1639 2100 1544 2568 R-squared 0.744 0.413 0.611 0.510 0.495 id fe yes yes yes yes yes year fe yes yes yes yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and. 10%, respectively. Heterogeneity analysis: Geographical location. Given China’s vast territory, cities show considerable regional differences in natural geographic elements such as topography, climate, and resource availability. Thus, this study investigates whether the impacts of low-carbon city pilot policies exhibit geographical diversity, categorizing the sample into eastern, central, and western regions. As shown in Table 10 , Column (3), which represents cities in the western region, shows a regression coefficient of 0.689, significant at the 10% level. However, the coefficients for cities in the eastern region (Column 1) and central region (Column 2) are insignificant, indicating that the positive effects of low-carbon city pilot policies on AI industry development are present only in the western region. This empirically illustrates the geographical heterogeneity in the effects of policy implementation, likely closely related to the industrial foundation and resource endowment characteristics of the western region. Given the later start of AI industry development in western regions, the marginal effects of policy incentives and resource input may be more pronounced, reflecting differentiated policy implementation effects against the backdrop of China’s unbalanced regional development. Table 10 Heterogeneity analysis: Geographical location & Industrial Dependence VARIABLES (1) (2) (3) (4) (5) LCPP 0.337 0.390 0.689* 0.411*** 0.493 (1.25) (1.44) (1.78) (2.22) (1.51) Controls yes yes yes yes yes Observations 1762 1162 1,189 2,577 1,356 R-squared 0.571 0.645 0.406 0.483 0.628 id fe yes yes yes yes yes year fe yes yes yes yes yes Notes: The standard errors are indicated in parentheses. ⁎⁎⁎ , ⁎⁎ , and ⁎ denote significance. levels at 1%, 5% and. 10%, respectively. Heterogeneity analysis: Industrial Dependence. Resource-based cities and old industrial bases that are overly dependent on traditional industrial development models, while relatively slow in fostering emerging industries such as AI, confront structural imbalances that contribute to their ‘regional collapse’ predicament. Following Zhang et al. ( 2024 ), we use the 2015 mean secondary industry proportion across all cities as a benchmark to categorize sample cities into ‘high secondary industry proportion’ and ‘low secondary industry proportion’ groups, as shown in Columns (4) and (5) of Table 8 , to investigate whether low-carbon city construction has a more significant impact on industry-dependent cities. The regression results indicate that low-carbon city pilot policies significantly enhance AI industry development in cities with higher industrial dependence, yet show no significant effect on cities with lower industrial dependence. This might be due to the policies encouraging new-old momentum conversion in industry-dependent cities, enabling a reasonable transformation and optimization of existing industrial infrastructure, technological accumulation, and talent reserves for the growth of emerging industries, thus accelerating AI industry development. Discussion Using panel data from 285 Chinese cities between 2007 and 2022, this study evaluated the impact of China’s LCPP program on the development of the urban AI industry through a staggered DID approach. The main results indicate that the implementation of China’s LCPP plan significantly positively influences the growth of the urban artificial intelligence industry. Additionally, the study conducted robustness tests using a placebo test, instrumental variable method, PSM-DID, replacing explained variables, altering clustering methods, adjusting the timing of LCPP occurrence, and controlling for other policy influences. The estimated results are consistent with the main findings, demonstrating the robustness of our conclusions. Additionally, we found that environmental regulation may drive the development of the artificial intelligence industry through two mechanisms: the first is the reconstruction of the energy consumption structure. When the price of traditional energy sources such as coal-fired power experiences a rigid increase due to policy tools like environmental taxes and quota restrictions, the large-scale utilization of green energy reshapes the electricity cost curve with its low marginal cost. This creates a price depression for low-carbon energy, attracting the agglomeration and growth of the artificial intelligence industry through low-cost electricity, predominantly driving the clustering of the downstream segments of the low-emission industrial chain. The second is the green technology innovation mechanism. Green technology innovation, prompted by environmental regulation, has notable positive externality characteristics. Its knowledge spillover effect can lower the threshold for technology adoption in related industries, thereby assisting the artificial intelligence industry in acquiring and implementing energy-saving technologies at a reduced cost, which, in turn, promotes its growth, particularly enhancing the development of the industry’s high-emission upstream segments and low-emission downstream segments. Finally, through heterogeneity tests, we found that the positive impact of the LCPP plan on the urban artificial intelligence industry is only significant in the western region, in third-tier cities and below, as well as in cities outside the Yangtze River Economic Belt. Moreover, the positive impact of the LCPP plan on the development of the artificial intelligence industry in heavily industrialized cities is more pronounced. This finding indicates that the role of the LCPP plan in fostering the development of the urban artificial intelligence industry is influenced by urban characteristics and does not apply uniformly across all cities. Theoretical implications. The theoretical implications of this study are as follows: First, existing studies primarily concentrate on the effects of environmental regulation policies on the transformation and upgrading of traditional industries, such as promoting technological innovation in enterprises by restricting high pollution emissions (Yang, Jahanger, and Hossain, 2023 ), or examining the influence of environmental regulation on enterprise productivity and competitiveness (Dagestani et al., 2023 ). This study broadens the research perspective to include strategic emerging industries and empirically tests the role of environmental regulation in advancing the development of the artificial intelligence sector, which not only addresses a gap in existing research but also enhances the theoretical understanding of the economic impacts of environmental regulation policies. Second, current research on the development of the artificial intelligence industry primarily focuses on technological advancements, talent aggregation, and capital investment (Rashid & Kausik, 2024 ; Zhang, 2023 ; Babina, T et al., 2024 ). However, the role of policy factors in industrial development cannot be overlooked. This study introduces an innovative perspective on environmental regulation policies. By examining the implementation effects and mechanisms of LCPP, it highlights the significant role of policy factors in promoting the development of the artificial intelligence industry. This suggests that when analyzing the driving forces behind the growth of the artificial intelligence sector, we should consider not only technical and market factors but also the influence of the policy environment and institutional design. Third, this study reveals the diverse impact of low-carbon policies on the development of urban artificial intelligence industries, providing empirical evidence for the implementation of targeted and effective policies. The study found that the positive effects of low-carbon policies on the artificial intelligence sector are more pronounced in the western regions, third-tier and smaller cities, areas outside the Yangtze River Economic Belt, and heavily industrialized cities. This finding illustrates that when implementing low-carbon policies to promote the growth of emerging industries, it is crucial to fully consider regional differences and urban characteristics and to adopt strategies tailored to local conditions. Policy implications. The findings of this study offer important insights for formulating low-carbon policies and promoting AI industry development. Specific policy implications include: First, governments should utilize low-carbon city pilot programs as an essential tool for advancing AI industry development. Drawing from China’s successful experiences, we recommend that governments: (1) Establish national-level low-carbon city pilot projects, select pilot cities through competitive evaluation, and provide appropriate policy support and financial incentives; (2) Create a policy framework that encourages the synergistic development of low-carbon initiatives and the AI industry, facilitating the innovative application of AI technologies in low-carbon areas such as energy management and environmental monitoring; (3) Develop a comprehensive evaluation system for low-carbon cities, incorporating the level of AI industry development into the evaluation criteria, and guiding cities to prioritize the cultivation of emerging industries during their low-carbon transition. Second, the regional heterogeneity analysis reveals significant variations in the effectiveness of low-carbon pilot policies across different regions. In China, when promoting low-carbon city construction, the development needs of less-developed regions and industrial transformation cities should take precedence. Specifically: (1) Develop differentiated support policies for low-carbon pilots, offering preferential funding and project layout assistance to less-developed regions; (2) Create specialized support plans for industrial transformation in traditional industrial cities, encouraging the deep integration of traditional industries with AI technologies through the establishment of industrial transformation funds; (3) Establish regional cooperation mechanisms to enable experience sharing and technical collaboration on low-carbon technologies and the AI industry between developed and less-developed regions. Third, based on the analysis of the policy impact mechanism, it is recommended to enhance the role of low-carbon policies in promoting the artificial intelligence industry from multiple perspectives: (1) It is suggested to deepen the market-oriented reform of energy prices, establish a composite pricing system, expand the peak-valley electricity price difference, and strengthen the guiding influence of price signals; (2) It is proposed to create a technology collaborative innovation consortium, establish a ‘dual carbon-AI’ national laboratory, overcome common technological barriers, and implement a rapid review system for green patents and compulsory licensing; (3) Implement a policy promotion strategy that varies in time and space, and develop corresponding policies based on the characteristics of different regions, such as focusing on smart grids in eastern coastal cities, laying out green computing infrastructure in central and western energy hubs, and promoting digitalization and carbon reduction in traditional manufacturing within the old industrial base in Northeast China. Declarations Author contributions author one: Data curation, Formal analysis, Investigation, Visualization, Writing—original draft, Writing—review & editing. author two: Formulate research questions, Data curation, Writing—review & editing. author three.: Supervision, Writing—review & editing. Competing interests The authors declare no competing interests. Data availability The datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request. Ethical approval The study is exempt from review by the Ethics Committee since it is classified as nonhuman subject research, thereby waiving the need for informed consent. Informed consent This article does not contain studies involving human participants engaged by any of the authors. Additional information Correspondence and requests for materials should be addressed to author one. References Arthur, W. B. (1989). Competing technologies, increasing returns, and lock-in by historical events. The economic journal, 99(394), 116-131. Aghion, P., Dechezleprêtre, A., Hemous, D., Martin, R., & Van Reenen, J. (2016). Carbon taxes, path dependency, and directed technical change: Evidence from the auto industry. Journal of Political Economy, 124(1), 1-51. Antoniou, F., & Mageiropoulos, T. (2024). Ranking the Barriers to the Energy Upgrading of Buildings Using the Best-Worst Method. Sustainability, 16(22), 10143. Bai, J., Bai, S., Chu, Y., Cui, Z., Dang, K., Deng, X., ... & Zhu, T. (2023). Qwen technical report. arxiv preprint arxiv:2309.16609. Babina, T., Fedyk, A., He, A., & Hodson, J. (2024). Artificial intelligence, firm growth, and product innovation. Journal of Financial Economics, 151(1), 103745. Beck, T., Levine, R., & Levkov, A. (2010). Big bad banks? The winners and losers from bank deregulation in the United States. The Journal of Finance, 65(5), 1637–1667. Bertrand, M., & Mullainathan, S. (2003). Enjoying the quiet life? Corporate governance and managerial preferences. Journal of Political Economy, 111(5), 1043–1075. Chen, H., Feng, L., & Sun, X. (2024). Beyond central-local relations: the introduction of a new perspective on China’s environmental governance model. Humanities and Social Sciences Communications, 11(1), 1-12. Chen, M. (2022). A study of low-carbon development, urban innovation and industrial structure upgrading in China. International Journal of Low-Carbon Technologies, 17(2), 185–195. Cyranoski, D. (2018). Chinese firms enter the battle for AI talent. Nature, 553(7688), 260–261. Chetty, R., Looney, A., & Kroft, K. (2009). Salience and taxation: Theory and evidence. American Economic Review, 99(4), 1145–1177. Chong, Z., Qin, C., & Ye, X. (2017). Environmental regulation and industrial structure change in China: Integrating spatial and social network analysis. Sustainability, 9(8), 1465. Dong, X., Song, S., & Zhu, H. (2011). Industrial structure and economic fluctuation: Evidence from China. The Social Science Journal, 48(3), 468–477. Dagestani, A. A., Shang, Y., Schneider, N., Cifuentes-Faura, J., & Zhao, X. (2023). Porter in China: A quasi-experimental view of market-based environmental regulation effects on firm performance. Energy Economics, 126(10), 106966. Dong, Y., Tian, J., & Wen, Q. (2022). Environmental regulation and outward foreign direct investment: Evidence from China. China Economic Review, 76(4), 101877. Dong, Z. Q., & Wang, H. (2019). The "local-neighborhood" green technology progress effect of environmental regulation. China Industrial Economy, (1), 100-118. Fan, F., & Zhang, X. (2021). Transformation effect of resource-based cities based on PSM-DID model: An empirical analysis from China. Environmental Impact Assessment Review, 91, 106648. Fahad, S., Bai, D., Liu, L., & Baloch, Z. A. (2022). Heterogeneous impacts of environmental regulation on foreign direct investment: Do environmental regulation affect FDI decisions? Environmental Science and Pollution Research, 29(4), 5092–5104. Gao, Y. (2022). Unleashing the mechanism among environmental regulation, artificial intelligence, and global value chain leaps: A roadmap toward digital revolution and environmental sustainability. Environmental Science and Pollution Research, 30(10), 28107–28117. Hong Kong Productivity Council and Hong Kong Institute of Economics and Business Strategy at HKU Business School. 2024. "Hong Kong AI Industry Development Study." Hong Kong: HKPC and HIEBS. https://www.hkpc.org/sites/default/files/2024-04/hkpc_hku_ai_industry_development_study_en.pdf. Accessed November 20 Haas, A., & Osland, L. (2014). Commuting, migration, housing and labour markets: Complex interactions. Urban Studies, 51(3), 463–476. Hong, M., Chen, S., & Zhang, K. (2021). Impact of the "Low-Carbon City Pilot" policy on energy intensity based on the empirical evidence of Chinese cities. Frontiers in Environmental Science, 9(7), Article 717737. Hu, Y., Dai, X., & Zhao, L. (2022). Digital finance, environmental regulation, and green technology innovation: An Empirical Study of 278 Cities in China. Sustainability, 14(14), 8652. International Energy Agency. 2023. "Emissions Grew in 2023, but Clean Energy Is Limiting the Growth - CO2 Emissions in 2023 - Analysis - IEA." Accessed November 20. https://www.iea.org/reports/co2-emissions-in-2023/emissions-grew-in-2023-but-clean-energy-is-limiting-the-growth Krugman, P. (1992). Geography and trade. MIT press. Li, C., Liang, F., Liang, Y., & Wang, Z. (2023). Low-carbon strategy, entrepreneurial activity, and industrial structure change: evidence from a quasi-natural experiment. Journal of Cleaner Production, 427, 139183. Li, S., Zheng, X., Liao, J., & Niu, J. (2024). Low-carbon city pilot policy and corporate environmental performance: Evidence from a quasi-natural experiment. International Review of Economics & Finance, 89, 1248-1266. Liu, Y., Peng, Y., Wang, W., Liu, S., & Yin, Q. (2024). Does the pilot zone for green finance reform and innovation policy improve urban green total factor productivity? The role of digitization and technological innovation. Journal of Cleaner Production, 471, 143365. Luqman, M., Rayner, P. J., & Gurney, K. R. (2023). On the impact of urbanisation on CO2 emissions. NPJ Urban Sustainability, 3(1), 6. Mi, T., & Li, T. (2024). Industrial Intelligence and Carbon Emission Reduction: Evidence from China's Manufacturing Industry. Sustainability (2071-1050), 16(15). Makiela, K., & Ouattara, B. (2018). Foreign direct investment and economic growth: Exploring the transmission channels. Economic Modelling, 72, 296-305. Pei, Y., Zhu, Y., Liu, S., Wang, X., & Cao, J. (2019). Environmental regulation and carbon emission: The mediation effect of technical efficiency. Journal of Cleaner Production, 236, 117599. Popp, D. (2006). International innovation and diffusion of air pollution control technologies: the effects of NOX and SO2 regulation in the US, Japan, and Germany. Journal of Environmental Economics and Management, 51(1), 46-71. Porter, M. E., & Linde, C. V. D. (1995). Toward a new conception of the environment-competitiveness relationship. Journal of economic perspectives, 9(4), 97-118. Rashid, A. B., & Kausik, A. K. (2024). AI revolutionizing industries worldwide: A comprehensive overview of its diverse applications. Hybrid Advances, 100277. Ren, X., Zhang, X., Yan, C., & Gozgor, G. (2022). Climate policy uncertainty and firm-level total factor productivity: Evidence from China. Energy Economics, 113, 106209. Ren, H., Gu, G., & Zhou, H. (2022). Assessing the low-carbon city pilot policy on carbon emission from consumption and production in China: how underlying mechanism and spatial spillover effect?. Environmental Science and Pollution Research, 29(47), 71958-71977. Si, X. (2024). A study on the impact of talent introduction policies on urban mobility: A literature review. Journal of Education, Humanities and Social Sciences, 39(November), 154–160. Sun, X., Song, Y., & Zhao, P. Y. (2022). The impact of artificial intelligence on the employment of heterogeneous labor force: From the perspective of labor supply. Economic Problem Exploration, (02), 171–190. Tao, Weiliang, Shimei Weng, Xueli Chen, Fawaz Baddar ALHussan, and Malin Song, “Artificial Intelligence-Driven Transformations in Low-Carbon Energy Structure: Evidence from China.” Energy Economics, 2024, 136 (August), 107719. Wu, L., Sun, L., Qi, P., Ren, X., & Sun, X. (2021). Energy endowment, industrial structure upgrading, and CO2 emissions in China: Revisiting resource curse in the context of carbon emissions. Resources Policy, 74, 102329. Wu, G., Sun, M., & Feng, Y. (2024). How does the new environmental protection law affect the environmental social responsibility of enterprises in Chinese heavily polluting industries?. Humanities and Social Sciences Communications, 11(1), 1-14. Wang, X., Zhou, J., & Maitlo, Q. (2022). Effect of green technology innovation on the upgrading of the manufacturing value chain: evidence from China. Frontiers in Environmental Science, 10, 994323. Xing, Z., Guo, J., Zhang, Z., Xue, T., Yang, M., & Wu, W. (2024). Research on the Impact of Environmental Inequality on Labor Mobility—A Study Based on the China General Social Survey (CGSS). Sustainability (2071-1050), 16(22). Yang, S., Jahanger, A., & Hossain, M. R. (2023). How effective has the low-carbon city pilot policy been as an environmental intervention in curbing pollution? Evidence from Chinese industrial enterprises. Energy Economics, 118, 106523. Yin, H., & Su, W. (2024). Industrial synergy agglomeration, urban innovation capacity, and advanced manufacturing development. Economies, 12(5), 117. Yu, X., & Wang, P. (2021). Economic effects analysis of environmental regulation policy in the process of industrial structure upgrading: Evidence from Chinese provincial panel data. Science of the Total Environment, 753, 142004. Zhang, L., Li, Y., Kung, C. C., Wu, B., & Zhang, C. (2023). Impact of new talent settlement policy on housing prices: Evidence from 70 large and medium-sized Chinese cities. PloS one, 18(3), e0280317. Zhang, L., Lin, G., Lyu, X., & Su, W. (2024). Suppression or promotion: research on the impact of industrial structure upgrading on urban economic resilience. Humanities and Social Sciences Communications, 11(1), 1-14. Zhang, Z. (2023). The impact of the artificial intelligence industry on the number and structure of employments in the digital economy environment. Technological Forecasting and Social Change, 197, 122881. Zhu, C., & Lee, C. C. (2022). The effects of low-carbon pilot policy on technological innovation: Evidence from prefecture-level data in China. Technological Forecasting and Social Change, 183, 121955. Zou, C., Huang, Y., Wu, S., & Hu, S. (2022). Does “low-carbon city” accelerate urban innovation? Evidence from China. Sustainable Cities and Society, 83, 103954. Zhou, Q., & Qi, Z. (2023). Urban economic resilience and human capital: An exploration of heterogeneity and mechanism in the context of spatial population mobility. Sustainable Cities and Society, 99, 104983. Zhou, Y. H., Yang, L., & Jiang, S. S. (2023). Binding carbon emission reduction and employment: An investigation based on changes in enterprise and regional labor force. Economic Research, 58(7), 104–120. Additional Declarations No competing interests reported. Cite Share Download PDF Status: Under Review Version 1 posted Editorial decision: Revision requested 25 Jun, 2025 Reviews received at journal 28 May, 2025 Reviewers agreed at journal 27 May, 2025 Reviewers agreed at journal 26 May, 2025 Reviewers agreed at journal 18 Apr, 2025 Reviews received at journal 18 Apr, 2025 Reviewers agreed at journal 18 Apr, 2025 Reviewers invited by journal 18 Apr, 2025 Editor invited by journal 17 Apr, 2025 Editor assigned by journal 17 Apr, 2025 Submission checks completed at journal 31 Mar, 2025 First submitted to journal 18 Mar, 2025 You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-6253945","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":444755886,"identity":"424f1426-8775-49dd-a0d9-efba5933e0b5","order_by":0,"name":"luyuan tang","email":"","orcid":"","institution":"Beijing Institute of Technology","correspondingAuthor":false,"prefix":"","firstName":"luyuan","middleName":"","lastName":"tang","suffix":""},{"id":444755887,"identity":"ee1d72bf-f374-422b-bffa-f866b5c72af4","order_by":1,"name":"shiyao xie","email":"","orcid":"","institution":"Peking University","correspondingAuthor":false,"prefix":"","firstName":"shiyao","middleName":"","lastName":"xie","suffix":""},{"id":444755888,"identity":"c8b0c4e5-8528-4c2a-a1ca-cfde14d353ec","order_by":2,"name":"yuan xu","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA+ElEQVRIie3PMUvEMBTA8RcCcXle11cQ+xVSCqdD4L5KDiFTv4NK4FzO3X4Lp3ONFupy0rWDg1k69xbp4GC7ivY6OuQPWcL7kReAUOgflnDuXNcrjCJbgR6v3BGS3om1L7bmLH6oDGg9g0CNWYbiRUmXLwHmEGZhSYgOwe0/P3wP54tGs+6Q/01OOBiiy3dk9v5JDotlcaN5XOwmX6lIYosc3nY0kPVjowU/nSBQsg1pUaKAvB3J9QzCuXQDQcjFSLQ8RlIrmL/dGiSqLqQ2lBZ7byf/kkR1V371arWqbet7pZLF69Vzd5ha7Ec0HHYzfz4UCoVCv/YNEJxTyvAjJswAAAAASUVORK5CYII=","orcid":"","institution":"Beijing Institute of Technology","correspondingAuthor":true,"prefix":"","firstName":"yuan","middleName":"","lastName":"xu","suffix":""}],"badges":[],"createdAt":"2025-03-18 14:08:15","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-6253945/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-6253945/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":81372150,"identity":"acc63354-42bc-4dd0-9539-9f2ee4033d21","added_by":"auto","created_at":"2025-04-25 10:42:56","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":99304,"visible":true,"origin":"","legend":"\u003cp\u003eParallel trend test\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-6253945/v1/6cc184339b4307add0f271c7.png"},{"id":81372151,"identity":"6f920ae4-e496-4a12-8f15-31002b5bd070","added_by":"auto","created_at":"2025-04-25 10:42:56","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":70398,"visible":true,"origin":"","legend":"\u003cp\u003ePlacebo test\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-6253945/v1/f039710dcaacd9684cff378c.png"},{"id":81373327,"identity":"51ca5ee6-4b41-48ed-884f-239d673521fe","added_by":"auto","created_at":"2025-04-25 10:58:58","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":1566762,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-6253945/v1/1a9cfc20-5cbd-4e9c-97ff-2d8aec03efde.pdf"}],"financialInterests":"No competing interests reported.","formattedTitle":"Are low-carbon cities and the development of the artificial intelligence industry contradictory?Evidence from China","fulltext":[{"header":"Introduction","content":"\u003cp\u003eClimate change has become one of humanity's most significant challenges in the 21st century. In this global environmental governance effort, cities, as the primary centers of human activities, play a crucial role in achieving global climate objectives. Recent statistics show that urban areas contribute approximately 70% of global energy-related CO₂ emissions, and this proportion is on the rise (Luqman, Rayner, and Gurney, \u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In China, urban carbon emissions account for 70–80% of total national emissions, with some studies estimating this figure could reach 85% by 2030 (Liguo Zhang et al., \u003cspan citationid=\"CR50\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). These statistics not only emphasize the vital role of cities in combating climate change but also highlight the significant challenges facing China, the world’s largest developing nation, in reaching its carbon neutrality targets.\u003c/p\u003e \u003cp\u003eTo advance the construction of ecological civilization and lead the way toward green, low-carbon development while achieving national greenhouse gas reduction targets, China’s National Development and Reform Commission introduced the Low-Carbon City Pilot Policy. Since launching the first batch of pilots in 2010, additional pilots were implemented in 2012 and 2017. This gradual expansion strategy for pilots reflects China’s unique approach to policy implementation. The selection of pilot cities comprehensively considered regional development stages, resource endowments, and differences in industrial structures, aiming to explore replicable regional low-carbon development models with diverse characteristics. In recent years, with the introduction of the ‘dual carbon’ goals and the execution of the Medium and Long-term Planning for Carbon Peak and Carbon Neutrality, low-carbon city construction has entered a new development phase, featuring an increasingly refined policy framework and strengthened implementation.\u003c/p\u003e \u003cp\u003eThe Low-Carbon City Pilot Policy, recognized as a significant innovation in environmental governance, has garnered extensive academic interest concerning its implementation effects and economic impacts. Existing research has highlighted the policy’s substantial influence across various dimensions: empirical studies have identified notable improvements in enterprise Total Factor Productivity (TFP) levels (X. Ren et al., \u003cspan citationid=\"CR38\" class=\"CitationRef\"\u003e2022\u003c/span\u003e); in terms of innovation-driven development, the policy has fostered technological advancement through optimized resource allocation and reduced financing constraints (Zhu and Lee, \u003cspan citationid=\"CR53\" class=\"CitationRef\"\u003e2022\u003c/span\u003e); regarding industrial structure, the policy has shown inhibitory effects on traditional high-carbon industries while simultaneously stimulating emerging sectors (C. Li et al., \u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). In enhancing energy efficiency, the policy mainly affects regional energy intensity through technological innovation rather than mechanisms for optimizing industrial structure (Hong, Chen, and Zhang, \u003cspan citationid=\"CR24\" class=\"CitationRef\"\u003e2021\u003c/span\u003e). Additional research has examined the policy’s influence on carbon intensity and per capita carbon emissions, including potential impact mechanisms (H. Ren, Gu, and Zhou, \u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). However, these studies have largely centered on traditional industries or general economic indicators, with relatively limited exploration of emerging strategic industries, particularly artificial intelligence (AI) sector.\u003c/p\u003e \u003cp\u003eThe AI industry, as the cornerstone of the new generation of information technology revolution, plays a vital role in the development of China’s digital economy. In recent years, China has actively issued policy documents like the ‘New Generation Artificial Intelligence Development Plan’ and the ‘Three-Year Action Plan for Promoting New Generation Artificial Intelligence Industry Development,’ elevating AI industry growth to a matter of national strategy. However, the AI industry demonstrates unique dual characteristics: on one hand, AI technology can optimize energy systems and enhance resource utilization efficiency, contributing to environmental governance; on the other hand, the extensive computing facilities and data center operations involved in industry growth generate substantial carbon footprints. Research shows that global energy-related carbon dioxide emissions rose by 1.1% in 2023, reaching a historic high of 37.4\u0026nbsp;billion tons (International Energy Agency, \u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This duality complicates the relationship between low-carbon city policies and AI industry development, revealing its complexity and nuance.\u003c/p\u003e \u003cp\u003eCurrent academic research on the relationship between artificial intelligence and carbon emissions is still in its early stages. While some scholars have demonstrated that AI applications can effectively promote carbon reduction (Tao et al., \u003cspan citationid=\"CR42\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), the influence of low-carbon city policies on the development of the AI industry has not been systematically examined. This research gap makes it challenging to ascertain whether environmental policies hinder the growth of emerging industries, how to reconcile emission reduction targets with the needs of industrial development, and whether policy design should take into account industry-specific characteristics.\u003c/p\u003e \u003cp\u003eThis study investigates the impact of low-carbon city policies on the AI industry, with significant theoretical and practical implications. On the theoretical level, the research extends the industrial impact theory of environmental regulation, enhances understanding of constraints on emerging industry development, and perfects the theoretical framework for coordinating environmental and industrial policies. On the policy level, the research findings provide empirical evidence for optimizing low-carbon city policy design, aiding in balancing environmental protection with industrial development, and offering references for formulating differentiated industry support policies. On the practical level, the findings will assist city administrators in better understanding policy impacts, guiding enterprises in their green transformation, and promoting the sustainable development of the AI industry. The innovation of this study lies in its pioneering systematic examination of the impact of environmental policies on strategic emerging industries, filling a gap in related research fields, and providing crucial guidance for facilitating China’s green, low-carbon transition and the development of its digital economy.\u003c/p\u003e \u003cp\u003eThe subsequent chapters are organized as follows: Section 2 introduces the policy background and presents the research hypotheses; Section 3 describes the research design; Section 4 reports the main results; Section 5 presents the mechanism analysis results; Section 6 introduces the heterogeneity analysis; and finally, Section 7 draws the research conclusions based on the previous analyses and provides theoretical implications and policy recommendations.\u003c/p\u003e\n\u003ch3\u003eResearch background and hypotheses development\u003c/h3\u003e\n\u003cp\u003e \u003cb\u003ePolicy background.\u003c/b\u003e The implementation of low-carbon policies represents China’s strategic choice in addressing rapidly increasing carbon emissions and promoting green transformation. In 2005, China became the world’s largest carbon emitter, with emissions reaching 6.098\u0026nbsp;billion tons, surpassing the United States’ 5.880\u0026nbsp;billion tons. Faced with growing international pressure for emission reductions and domestic environmental constraints, China urgently needed to transform its development model. At the 2009 Copenhagen Climate Conference, China pledged to reduce its carbon intensity by 40–45% by 2020 compared to 2005 levels, marking China’s transition from an extensive to a low-carbon development model.\u003c/p\u003e \u003cp\u003eIn this context, China launched three rounds of low-carbon city pilot programs in July 2010, November 2012, and February 2017. The first round consisted of five provinces (Guangdong, Liaoning, Hubei, Shaanxi, and Yunnan) and eight cities (Tianjin, Chongqing, Shenzhen, Xiamen, Hangzhou, Nanchang, Guiyang, and Baoding). The second round included 28 cities, such as Beijing, Shanghai, and Qinhuangdao, along with Hainan Province. The third round, building on the successes of the previous two, covered 45 cities (districts and counties). Due to the extensive scope of the third round, pilot regions were required to develop replicable and scalable experiences, with successful practices gradually being promoted nationwide by 2020. Therefore, this study considers 2010, 2012, and 2017 as the implementation time points for the low-carbon city pilot policy.\u003c/p\u003e \u003cp\u003e \u003cb\u003eHypotheses development: The impact of Low-carbon city on artificial intelligence industry development.\u003c/b\u003e Existing literature presents two competing explanations for the relationship between environmental regulation and industrial upgrading: the traditional school argues that environmental regulation fosters the upgrading of industrial structure through the factor relocation effect (Zhaohui Chong et al., \u003cspan citationid=\"CR12\" class=\"CitationRef\"\u003e2017\u003c/span\u003e), while the revisionist school highlights that the institutional environment may create an upgrading barrier effect (Antoniou, F., \u0026amp; Mageiropoulos, T., \u003cspan citationid=\"CR3\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). This theoretical divergence offers a distinct perspective on understanding the AI industry effect of the low-carbon city pilot policy.\u003c/p\u003e \u003cp\u003eFrom the theoretical perspective of the revisionist school, within a specific institutional context, environmental regulation may create a ‘compliance burden.’ To fulfill short-term environmental compliance requirements, enterprises allocate limited resources to end-of-pipe governance instead of pursuing fundamental technological innovation (Wu, G et al., \u003cspan citationid=\"CR44\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), which leads to insufficient motivation for innovation. Simultaneously, under the cost pressures of environmental regulation, enterprises may prioritize established technology paths over cutting-edge exploration, thereby resulting in a ‘low-end lock-in’ effect (Wang, X et al., \u003cspan citationid=\"CR45\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). This structural constraint presents a potential hindrance to the development of emerging industries.\u003c/p\u003e \u003cp\u003eHowever, China’s institutional context may undermine the theoretical effects outlined above. As a large developing country, China often takes a gradual reform approach to environmental regulation, using a ‘pilot first’ strategy that grants local governments greater autonomy in policy implementation (Chen, H et al., \u003cspan citationid=\"CR8\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). In this framework, environmental regulation generally pairs with supporting industrial policies, creating a dual policy combination of ‘environmental constraints + innovation incentives’. As a comprehensive reform initiative, the low-carbon city pilot not only establishes carbon emission constraints but also offers innovation incentives such as fiscal subsidies, tax breaks, and technical support, which effectively mitigates the potential innovation-inhibiting effects that might arise from basic environmental regulations. Furthermore, under the performance evaluation system, local governments in China often leverage environmental regulation as a catalyst for industrial upgrading, facilitating the concentration of high-tech industries in their regions through a mix of investment promotion and industrial initiatives policies.\u003c/p\u003e \u003cp\u003eBased on the theoretical analysis and considerations of real-world situations outlined above, this study proposes the following core hypothesis:\u003c/p\u003e \u003cp\u003eH1: The low-carbon city pilot policy can foster the development of the artificial intelligence industry.\u003c/p\u003e \u003cp\u003eThe effect of the low-carbon city pilot policy on the artificial intelligence industry primarily operates through the cost transmission mechanism, shaped by environmental regulation constraints, and the knowledge spillover mechanism associated with technological externalities. According to the new economic geography theory (Krugman, \u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e1992\u003c/span\u003e) and technology spillover theory, environmental regulation can enhance industrial development advantages through the dual mechanisms of optimizing energy structures and promoting green technology innovation.\u003c/p\u003e \u003cp\u003e \u003cb\u003eHypotheses development: The mediating role of energy consumption structure.\u003c/b\u003e As the central subject of carbon emissions within the urban economic system, the low-carbon transformation of enterprises directly influences the effectiveness of urban environmental governance. From a cost-based theoretical perspective, environmental regulatory policies have raised the prices of traditional energy sources, encouraging artificial intelligence companies to seek lower-cost energy alternatives. This shift optimizes their green energy consumption structure and significantly enhances the geographical clustering of the industry. Specifically, on one hand, the low-carbon city pilot policy has notably increased the costs associated with traditional energy sources like coal-fired power through mechanisms such as carbon emission rights trading, environmental taxes, and differentiated electricity pricing. On the other hand, the policy has spurred the growth of renewable energy capacity and technological advancements through the green power quota system and priority dispatch mechanism, leading to a continuous reduction in the marginal costs of clean energy sources like wind and solar power. As the supply of green energy reaches a critical scale, its low marginal cost characteristics will reshape the regional power market dynamics, resulting in a competitive decline in low-carbon energy prices. This structural change exerts a significant gravitational influence on the artificial intelligence industry, which has high electricity demands. When pilot areas develop an electricity pricing advantage, both upstream and downstream enterprises within the artificial intelligence industry chain will naturally form spatial clusters based on cost efficiency and operational stability. Accordingly, this paper proposes:\u003c/p\u003e \u003cp\u003eH2: The low-carbon city pilot policy attracts the agglomeration and development of the artificial intelligence industry by changing the energy consumption structure.\u003c/p\u003e \u003cp\u003e \u003cb\u003eHypotheses development\u003c/b\u003e: \u003cb\u003eThe mediating role of green technology innovation.\u003c/b\u003e In addition to influencing industrial agglomeration by altering the energy consumption structure, environmental regulation can also impact the path of technology diffusion by stimulating innovation activities.\u003c/p\u003e \u003cp\u003eFrom the perspective of innovation compensation, moderate environmental regulation can encourage enterprises to pursue green technology innovation, and its knowledge spillover effect can lower the technology adoption threshold for related industries, thereby facilitating the industry's growth. Specifically, there are notable differences between green technology innovation and traditional technology innovation, as green innovation often exhibits stronger universality and positive externality characteristics (Hu Y et al., \u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e2022\u003c/span\u003e). For instance, technologies for improving energy efficiency and carbon capture and storage have application principles and solutions that can be transferred across industries, making green innovation more likely to produce knowledge spillover effects. Conversely, the artificial intelligence industry is significantly technology-intensive, requiring continuous integration and application of cross-domain technologies. Green technology innovation triggered by environmental regulation, particularly breakthroughs in energy efficiency optimization, smart grids, and distributed energy management, aligns well with the technological needs of the artificial intelligence industry. This technological connection allows the knowledge spillover of green innovation to reduce the technology adoption threshold for artificial intelligence companies. From the perspective of technological economics, reducing technology adoption costs directly impacts the industry's development trajectory. The knowledge spillover effect of green technology innovation not only lowers the R\u0026amp;D costs for artificial intelligence companies but also provides them with new application scenarios and business models, further promoting industrial development. Based on this, this paper proposes:\u003c/p\u003e \u003cp\u003eH3: The low-carbon city pilot policy triggers green technology innovation and enables the growth of the artificial intelligence industry.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e\u003c/p\u003e \u003cp\u003e\u003c/p\u003e \u003c/div\u003e"},{"header":"Research design","content":"\u003cp\u003eThe definitions and the descriptive statistics of the variables utilized in this study are given in Table\u0026nbsp;\u003cspan refid=\"Tab1\" class=\"InternalRef\"\u003e1\u003c/span\u003e.\u003c/p\u003e\u003cdiv class=\"gridtable\"\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e\u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e\u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e\u003ctable float=\"Yes\" id=\"Tab1\" border=\"1\"\u003e\u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 1\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eDescriptive statistics\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e\u003ccolgroup cols=\"6\"\u003e\u003c/colgroup\u003e\u003cthead\u003e\u003ctr\u003e\u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eN\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eSd\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eMin\u003c/p\u003e \u003c/th\u003e\u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003eMax\u003c/p\u003e \u003c/th\u003e\u003c/tr\u003e\u003c/thead\u003e\u003ctbody\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e4.576\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.921\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.693\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e9.707\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.300\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.458\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e5.872\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.672\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.822\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e7.116\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDens\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.271\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.461\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.00527\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e3.026\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGdp\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e10.58\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.675\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e8.919\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e12.05\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFina\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e2.383\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e1.140\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.896\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e6.760\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpen\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.180\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.283\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.00165\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1.657\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEdu\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.0185\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.0236\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.000629\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.117\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003ctr\u003e\u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInf\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c3\"\u003e \u003cp\u003e0.815\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"char\" char=\".\" colname=\"c4\"\u003e \u003cp\u003e0.374\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e\u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2.245\u003c/p\u003e \u003c/td\u003e\u003c/tr\u003e\u003c/tbody\u003e\u003c/table\u003e\u003c/div\u003e\u003cp\u003e \u003cb\u003eDependent Variable: Development of artificial intelligence industry (DAII).\u003c/b\u003e To measure the development levels of the AI industry, following the methodology of the Hong Kong Productivity Council (2024) and Sun et al. (\u003cspan citationid=\"CR41\" class=\"CitationRef\"\u003e2022\u003c/span\u003e), we utilized Python web crawling technology to perform fuzzy keyword matching queries on the ‘business scope’ and ‘company name’ fields in the Qichacha database. We employed keywords related to AI applications, including ‘artificial intelligence,’ ‘cloud,’ ‘data,’ ‘IoT,’ and ‘machine learning.’ The data was aggregated by year and region to create panel data of AI enterprises in Chinese cities from 2007 to 2022. The natural logarithm of the number of AI enterprises in each city-year observation was employed to measure AI industry development (DAII).\u003c/p\u003e\u003cp\u003e \u003cb\u003eIndependent Variable: Low-carbon city pilot policy (LCPP).\u003c/b\u003e We employed the staggered Difference-in-Differences (DID) method to examine the implementation effects of the low-carbon city pilot policy (Popp, \u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e2006\u003c/span\u003e; Pei et al., \u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e2019\u003c/span\u003e). When the pilot region was a province, all cities within that province were regarded as pilot cities. If the same city appeared in different batches of pilot lists, the earliest batch was utilized to establish the pilot start time. Accordingly, we constructed a policy dummy variable (LCPP) as follows:\u003c/p\u003e\u003cdiv id=\"Equ1\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ1\" name=\"EquationSource\"\u003e\n$$\\:LCP{P}_{i,t}=Trea{t}_{i}\\times\\:Tim{e}_{t}\\:\\:\\:\\:\\:\\:\\:\\:\\:$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e1\u003c/div\u003e\u003c/div\u003e\u003cp\u003eWhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Trea{t}_{i}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Tim{e}_{t}\\)\u003c/span\u003e\u003c/span\u003e are dummy variables, \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003e represents city \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e represents year \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Trea{t}_{i}\\)\u003c/span\u003e\u003c/span\u003e equals 1 when the city is the pilot city, and 0 otherwise. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\text{T}\\text{i}\\text{m}{\\text{e}}_{\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e takes the value of 1 when the policy is actually, and 0 otherwise.\u003c/p\u003e\u003cp\u003e \u003cb\u003eControl variables.\u003c/b\u003e Following previous research (S. Li et al., \u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), we controlled for several city characteristics that might influence the development of the AI industry. These variables include: (1) the natural logarithm of year-end total population, representing population size (Size); (2) GDP divided by administrative land area, representing urban economic density (Dens); (3) the natural logarithm of GDP per capita, representing economic development level (Gdp); (4) year-end financial institution loan balance divided by GDP, representing the degree of financial development (Fina); (5) total import-export volume divided by GDP, representing the degree of openness (Open); (6) the number of students in higher education institutions per year-end total population, representing education level (Edu); and (7) fixed asset investment divided by GDP, representing infrastructure level (Inf).\u003c/p\u003e\u003cp\u003e \u003cb\u003eModel settings.\u003c/b\u003e The DID method is the most widely recognized model in policy evaluation literature, as it significantly mitigates omitted variable bias and accurately identifies exogenous policy shocks. Considering the three batches of low-carbon city pilots, we adopted a ‘staggered DID’ approach to study the impact of low-carbon city construction on AI industry development.\u003c/p\u003e\u003cdiv id=\"Equ2\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ2\" name=\"EquationSource\"\u003e\n$$\\:DAI{I}_{i,t}=\\alpha\\:+\\beta\\:LCP{P}_{i,t}+\\lambda\\:\\sum\\:Control{s}_{i,t}+{\\gamma\\:}_{t}+{\\mu\\:}_{i}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e2\u003c/div\u003e\u003c/div\u003e\u003cp\u003eWhere \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:i\\)\u003c/span\u003e\u003c/span\u003eand \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:t\\)\u003c/span\u003e\u003c/span\u003e denote city and year respectively.\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:DAI{I}_{i,t}\\)\u003c/span\u003e\u003c/span\u003edenotes development of artificial intelligence industry. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:LCP{P}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e is a policy dummy variable, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\beta\\:\\)\u003c/span\u003e\u003c/span\u003e shows how the LCPP affects DAII. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:Controls\\:\\)\u003c/span\u003e\u003c/span\u003edenotes a series of city characteristics controlled for in this paper. \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\gamma\\:}_{t}\\)\u003c/span\u003e\u003c/span\u003e and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\mu\\:}_{i}\\)\u003c/span\u003e\u003c/span\u003e denote the fixed effect of year and city respectively, and \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\epsilon\\:}_{i,t}\\)\u003c/span\u003e\u003c/span\u003e is a random interference item.\u003c/p\u003e\u003cp\u003e \u003cb\u003eData sources.\u003c/b\u003e The study covers data from 2007 to 2022, including 285 prefecture-level and higher cities in China (excluding those with significant data gaps). Data on AI industry development was obtained from the Chinese enterprise credit information platform, ‘Qichacha.’ The low-carbon city pilot list was sourced from relevant documents of the National Development and Reform Commission. Additional data came from the Express Professional Superior (EPS) platform and the ‘China City Statistical Yearbook.’ Linear interpolation was employed to fill in minor missing data in some cities.\u003c/p\u003e"},{"header":"Results and discussion","content":"\u003cp\u003e \u003cb\u003eMain results.\u003c/b\u003e Based on the aforementioned benchmark model, this section examines the quantitative impact of implementing the low-carbon city pilot policy on the development of the urban artificial intelligence industry. The estimated results are presented in Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The coefficient of the LCPP is the focus of our analysis. In column (1) of Table\u0026nbsp;\u003cspan refid=\"Tab2\" class=\"InternalRef\"\u003e2\u003c/span\u003e, without the addition of control variables and fixed effects, the LCPP coefficient indicates a positive effect at the 1% significance level, which preliminarily verifies the policy's effectiveness. Subsequently, after introducing both the city-year two-way fixed effects and all control variables in columns (2) and (3), the coefficient of the core explanatory variable still holds 1% statistical significance, demonstrating the robustness of the main results. For each 1 percentage point increase in the LCPP policy, the DAII increases by 1.39 percentage points. This quantitative finding effectively confirms research hypothesis H1, which states that the low-carbon city pilot policy can promote the development of the urban artificial intelligence industry.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab2\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 2\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMain results\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1.391***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.699***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.156***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(23.75)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(13.91)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(6.59)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.038***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(65.09)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDens\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.250***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(8.59)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGdp\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.323***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(52.10)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.297***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(21.17)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.465***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(8.96)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEdu\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e7.144***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(10.91)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInf\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.165***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(5.38)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eConstant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4.159***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.212***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-17.565***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(129.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(27.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(-61.76)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.487\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.885\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eno\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003elevels at 1%, 5%. and. 10%, respectively.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eParallel trend test.\u003c/b\u003e The prerequisite for the DID method is that both the experimental and control groups meet the parallel trends assumption prior to the policy implementation. This means there are no systematic differences in AI industry development between pilot and non-pilot cities before the introduction of the low-carbon city pilot policy. Eq.\u0026nbsp;(\u003cspan refid=\"Equ3\" class=\"InternalRef\"\u003e3\u003c/span\u003e) is structured to examine the dynamic trends of digital government policy (Beck et al., \u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e2010\u003c/span\u003e; Bertrand and Mullainathan, \u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e2003\u003c/span\u003e), as follows:\u003cdiv id=\"Equ3\" class=\"Equation\"\u003e\u003cdiv format=\"TEX\" class=\"mathdisplay\" id=\"FileID_Equ3\" name=\"EquationSource\"\u003e\n$$\\:DAI{I}_{i,t}={\\beta\\:}_{0}+{\\sum\\:}_{k=-5}^{10}{\\beta\\:}_{k}LCP{P}_{i,t}^{k}+{\\beta\\:}_{2}Control{s}_{i,t}++{\\gamma\\:}_{t}+{\\mu\\:}_{i}+{\\epsilon\\:}_{i,t}$$\u003c/div\u003e\u003cdiv class=\"EquationNumber\"\u003e3\u003c/div\u003e\u003c/div\u003e\u003c/p\u003e \u003cp\u003eTo provide a more intuitive view of the results from the parallel trend test, we plotted the estimated values of the coefficients along with their 95% confidence intervals. The results of the parallel trend test are illustrated in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003e below. Prior to the policy's implementation, the regression coefficient for LCCP did not demonstrate statistical significance, indicating that there was no notable difference in carbon emission efficiency between pilot cities and non-pilot cities before the policy intervention. Following the policy implementation, the LCCP coefficient began to show a continuous upward trend. This trend aligns with the main regression results, suggesting that the policy may positively influence carbon emission efficiency. The increase in the coefficient suggests that the improvement in carbon emission efficiency is notably greater in the intervened cities compared to the non-intervened cities. Notably, one year after the policy was enacted, the coefficient achieved statistical significance, indicating a potential one-year lag effect of LCCP. This lag effect may arise from the delay between policy planning and execution, or it could reflect the time cities need to adjust and respond to policy requirements.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003ePlacebo test.\u003c/b\u003e Although this study has eliminated as many confounding factors as possible that might influence the results, such as including city and time fixed effects in the regression, these measures alone are insufficient to overcome the problems arising from omitted variable bias. Therefore, we further conduct robustness checks through placebo tests. Following Chetty et al. (\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e2009\u003c/span\u003e), we randomly select a group of cities as pseudo-pilot cities, reconstruct the policy dummy variable, and obtain estimation coefficients. The results of 500 counterfactual estimations are shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003e. The pseudo-regression coefficients are concentrated around zero, approximating a normal distribution, while the baseline regression estimation coefficient falls outside this distribution. Therefore, we can claim that our baseline estimation results in this article are not swayed by unobservable factors, stipulating that the findings of this article are replicable.\u003c/p\u003e \u003cp\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eEndogeneity analysis.\u003c/b\u003e Reverse causality is one of the significant reasons for the endogeneity problem. Specifically, the low-carbon city pilot policy is not entirely exogenous, and the developmental needs of the regional artificial intelligence industry may in turn promote the implementation of the policy, creating a two-way causal relationship. Therefore, this paper employs the instrumental variable method to address the potential endogeneity issue in the benchmark regression. This paper constructs two sets of geographical-historical instrumental variables. First, this paper follows the approach of Zhou (2024) to manually collect government work reports from 285 prefecture-level cities from 2008 to 2022, segment the text of the reports, count the frequency of words related to environmental regulation and environmental protection, and calculate the proportion of these words in relation to the total word frequency of the full text of the government report as an instrumental variable (IV1). On one hand, the frequency of words associated with environmental regulation and protection in the government work report indicates the level of attention local governments give to environmental issues. This level of attention is to some extent independent of the implementation effect of the low-carbon city pilot policy. Therefore, this instrumental variable can be regarded as meeting the exclusivity requirement. On the other hand, this instrumental variable is highly correlated with the implementation of the low-carbon city pilot policy, as greater attention from local governments to environmental matters increases the likelihood of actively promoting the low-carbon city pilot policy, thus fulfilling the correlation requirement and qualifying as a suitable instrumental variable. The low-carbon city pilot policy stimulates green technology innovation and facilitates the development of the artificial intelligence industry.\u003c/p\u003e \u003cp\u003eSecond, we refer to the research of Dong Zhiqinghua and Wang Hui (2019) and construct the industrial pollution stock index before the policy implementation as an instrumental variable (IV2): ① Select six major pollutants such as SO₂, smoke, and COD; ② Use the entropy method to determine the weights: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{W}}_{\\text{j}\\text{t}}=\\frac{1-{\\text{E}}_{\\text{j}\\text{t}}}{{\\sum\\:}_{\\text{k}=1}^{6}(1-{\\text{E}}_{\\text{k}\\text{t}})}\\)\u003c/span\u003e\u003c/span\u003e, where \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{E}}_{\\text{j}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e=\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:-{\\sum\\:}_{\\text{i}=1}^{\\text{N}}\\frac{{\\text{E}}_{\\text{i}\\text{j}\\text{t}}}{{\\text{E}}_{\\text{j}\\text{t}}}\\text{l}\\text{n}\\frac{{\\text{E}}_{\\text{i}\\text{j}\\text{t}}}{{\\text{E}}_{\\text{j}\\text{t}}}\\)\u003c/span\u003e\u003c/span\u003e;③ Calculate the standardized pollution intensity in batches: \u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{P}\\text{I}}_{\\text{i}\\text{t}}^{\\text{b}}\\)\u003c/span\u003e\u003c/span\u003e=\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\sum\\:_{\\text{j}=1}^{6}(\\frac{{\\text{E}}_{\\text{i}\\text{j}\\text{t}}/{\\text{G}\\text{D}\\text{P}}_{\\text{i}\\text{t}}}{\\frac{1}{{\\text{N}}_{\\text{b}}}{\\sum\\:}_{\\text{i}=1}^{{\\text{N}}_{\\text{b}}}{\\text{E}}_{\\text{i}\\text{j}\\text{t}}/{\\text{G}\\text{D}\\text{P}}_{\\text{i}\\text{t}}}\\)\u003c/span\u003e\u003c/span\u003e)\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:\\times\\:{\\text{W}}_{\\text{j}\\text{t}}\\)\u003c/span\u003e\u003c/span\u003e(representing the number of candidate cities in batch b).\u003c/p\u003e \u003cp\u003eOn one hand, from the correlation perspective, the stock of industrial pollution prior to the policy's implementation can reflect the severity of environmental issues in a region. This severity often influences local government\u0026rsquo;s attention to and execution of the low-carbon city pilot policy. Therefore, this instrumental variable is strongly correlated with the low-carbon city pilot policy. On the other hand, from the exogeneity perspective, the industrial pollution stock before the policy's implementation is established prior to the policy itself and is unrelated to the economic effects that occur post-implementation. Therefore, this instrumental variable satisfies the exogeneity requirement, indicating its effectiveness.\u003c/p\u003e \u003cp\u003eThis paper employs the two instrumental variables mentioned above for two-stage estimation, with specific regression results presented in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e. We report the estimation outcomes of IV1 and IV2 in columns (1) through (4). The results indicate that in the first stage, the first instrumental variable significantly positively impacts LCPP at the 5% level, while the second instrumental variable shows a significant positive impact on LCPP at the 1% level, confirming our analysis. In the second stage, both instrumental variables are introduced individually, and the estimated coefficients of LCPP on DAII are significantly positive at the 1% level. The model passes both the unidentification test and weak identification test, suggesting that the two instrumental variables developed in this paper are sufficiently exogenous and that there is no weak instrumental variable issue. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab3\" class=\"InternalRef\"\u003e3\u003c/span\u003e, the estimation results of this paper remain unaffected by the endogeneity problem, maintaining robustness.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab3\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 3\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eEndogeniety analysis\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\" morerows=\"2\" rowspan=\"3\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003ePhase 1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003ePhase 2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003ePhase 1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003ePhase 2\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eIV1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eIV1\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eIV2\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003eIV2\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e3.570***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4.009***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(4.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4.50)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIV1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.151**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(2.08)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eIV2\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.164***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(4.82)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.107\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.112\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.111\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKleibergen-Paap rk LM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e70.4547***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e69.5295***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eKleibergen-Paap rk Wald F\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e69.110\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e71.820\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance levels\u003c/p\u003e \u003cp\u003eat 1%, 5%. and. 10%, respectively.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePropensity score matching (PSM)-DID.\u003c/b\u003e To control for potential sample selection bias, we employ the PSM-DID method for re-estimation (Fei \u0026amp; Zhang, 2021). This study uses nearest neighbor matching for verification. Using control variables as matching variables shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, before matching, the ATT (Average Treatment Effect on the Treated) is 5.5300 and significant at the 1% level (absolute t-value greater than 2.58); after matching, the ATT is 0.8542, also significant at the 1% level. In the balance test shown in Table\u0026nbsp;\u003cspan refid=\"Tab4\" class=\"InternalRef\"\u003e4\u003c/span\u003e, the %bias of matching variables is less than 10% and notably smaller than the pre-matching %bias, indicating no significant differences between the control and experimental groups after PSM.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab4\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 4\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eBalancing test results of PSM\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"9\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c7\" colnum=\"7\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c8\" colnum=\"8\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c9\" colnum=\"9\"\u003e\u003c/div\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eVariables\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eUnmatched\u003c/p\u003e \u003cp\u003eMatched\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c4\" namest=\"c3\"\u003e \u003cp\u003eMean\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c6\" namest=\"c5\"\u003e \u003cp\u003e%Reduction\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colspan=\"2\" nameend=\"c8\" namest=\"c7\"\u003e \u003cp\u003et-test\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003eV(T)/V(C)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eTreated\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eControl\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e%Bias\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e[Bias]\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003et\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003ep\u0026thinsp;\u0026gt;\u0026thinsp;t\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.889\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.866\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e72.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.00\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.316\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.94\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e5.883\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e5.877\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.9\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.22\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.825\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.77*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDens\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.542\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.196\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e90.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14.57\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e19.09*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.448\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.413\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e3.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e1.31\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.191\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.56*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGdp\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.875\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.448\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e66.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e99.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e19.95\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.75*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e10.863\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e10.865\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-0.09\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.925\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.92\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.805\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.223\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e45.6\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e99.4\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e14.87\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.67*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e2.783\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2.789\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e-0.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e-0.05\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.959\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.43*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.242\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.161\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e24.8\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e99.7\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e8.04\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.60*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.232\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.231\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.02\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.986\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e0.94\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eEdu\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.016\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e37.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e65.0\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e12.47\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.000\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e2.05*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.025\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.022\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e13.1\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e3.20\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.001\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.48*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eInf\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eU\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.813\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.808\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\" morerows=\"1\" rowspan=\"2\"\u003e \u003cp\u003e61.3\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.39\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.696\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.20*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eM\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.813\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.811\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.5\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c7\"\u003e \u003cp\u003e0.12\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c8\"\u003e \u003cp\u003e0.901\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c9\"\u003e \u003cp\u003e1.14*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"9\"\u003eNotes: The standard errors are indicated in parentheses. ⁎⁎⁎, ⁎⁎, and ⁎ denote significance levels at 1%, 5% and 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e presents the PSM-DID estimation results for DAII, with significantly positive estimation coefficients, validating the robustness of our main results.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab5\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 5\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003ePSM-DID estimation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"4\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"char\" char=\".\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003eDAII\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.083***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eEdu\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-8.124***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(5.43)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(-4.19)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eSize\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e2.450***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eInf\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1.006***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(4.42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(5.21)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eDens\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e5.096***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eConstant\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-10.523***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(16.54)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(-2.62)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGdp\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e-0.787***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,420\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(-5.15)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.987\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eFina\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.044\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(0.76)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eOpen\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e0.616**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"char\" char=\".\" colname=\"c2\"\u003e \u003cp\u003e(2.00)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"4\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cdiv class=\"BlockQuote\"\u003e \u003cp\u003elevels at 1%, 5% and 10%, respectively.\u003c/p\u003e \u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eSubstitution of explained variable.\u003c/b\u003e From the sample observation, it is found that the business scope provided at the time of enterprise registration is relatively broad, leading to excessive granularity when directly capturing and identifying artificial intelligence enterprises based on their name or business scope. Therefore, to more accurately define and analyze the artificial intelligence industry, we utilize the national standard classification code to categorize the ten industries of software development, technology promotion services, other science and technology promotion services, other information technology services, engineering and technology research and experimental development, information technology consulting services, consulting and investigation, Internet information services, information system integration, and Internet of Things technology services as core artificial intelligence enterprises, while classifying the remaining industries as marginal enterprises. This study employs the number of regional core artificial intelligence enterprises to represent the scale of the regional artificial intelligence industry as an alternative dependent variable. The regression results from Table\u0026nbsp;\u003cspan refid=\"Tab5\" class=\"InternalRef\"\u003e5\u003c/span\u003e (1) indicate that the impact coefficient is 0.262, which is significant at the 1% level, supporting the hypothesis H1.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab6\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 6\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eRobustness tests: other methods\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"5\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.262***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.495***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.551***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.316***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(3.42)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2.68)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(5.71)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(5.73)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,259\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4,259\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.497\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.718\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.509\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"5\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance. levels at 1%, 5% and 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eChanging clustering method.\u003c/b\u003e To account for clustering of standard errors, we clustered them at the province and year levels for robustness checks. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e Column (2), the estimated coefficient for low-carbon city construction remains significant at the 1% level, confirming the robustness of the baseline regression results.\u003c/p\u003e \u003cp\u003e \u003cb\u003eConsidering implementing lag one-stage.\u003c/b\u003e Given that the impact of the LCPP intervention might not be immediate, we lagged the core regressor (LCCP) by one period. We applied the same method to all control variables to address simultaneous equation bias. As shown in Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e, Column (3), the LCPP policy demonstrates a positive promotional effect on the development of the urban AI industry at the 1% significance level. This finding is consistent with the baseline regression results, further reinforcing our previous conclusions.\u003c/p\u003e \u003cp\u003e \u003cb\u003eEliminating the influence of COVID-19.\u003c/b\u003e To ensure the accuracy of these findings, it was essential to eliminate the impact of COVID-19 in 2020, as it could significantly affect data distribution, trends, and outliers. Neglecting to account for this influence could lead to biased or inaccurate results. Therefore, this study excluded data from 2020 to remove the effects of COVID-19. Column (4) of Table\u0026nbsp;\u003cspan refid=\"Tab6\" class=\"InternalRef\"\u003e6\u003c/span\u003e shows that after controlling for COVID-19 effects, the LCCP policy estimation remains significant at the 1% level, demonstrating the robustness of the baseline regression model.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMechanism analysis:Energy consumption structure.\u003c/b\u003e According to the theoretical analysis presented in the previous article, the low-carbon city pilot policy may facilitate the agglomeration development of the artificial intelligence industry by optimizing the energy consumption structure. To verify this mechanism, the proportion of regional coal consumption\u0026mdash;directly related to the energy consumption structure\u0026mdash;is used here to represent the regional energy consumption structure for testing hypothesis 2. Column (1) of Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e displays the estimated results of the impact of the low-carbon city pilot policy on the energy consumption structure. The results reveal that the coefficient estimate of the LCPP is -0.112, and it passes the significance test at the 1% level. This indicates that, under cost theory, the low-carbon city pilot policy significantly reduces the intensity of regional use of traditional energy. This paper further examines the impact of energy consumption structure on the development of the artificial intelligence industry. The results in column (2) indicate that the coefficient estimate of energy consumption structure is -0.037, which also passes the significance test at the 1% level. This suggests that the low-carbon city pilot policy can promote the development of the regional artificial intelligence industry by decreasing the intensity of regional consumption of traditional energy, such as coal.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab7\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 7\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMechanism analysis: energy consumption structure\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(5)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVARIABLES\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eNY\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eDAII\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eUpstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eMidstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eDownstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCCP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e-0.112***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(-3.07)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eNY\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e-0.037***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e-0.082\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.032*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e-0.024***\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(-2.64)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(-1.14)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(1.62)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(-1.68)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.883\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.367\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.349\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.139\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.171\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance. levels at 1%, 5% and 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003eThis paper also examines the impact of the energy consumption structure on various aspects of the artificial intelligence industry chain. Referring to the industry-level division paradigm officially proposed by Qichacha, the artificial intelligence industry chain is divided into three segments: the basic layer (upstream link), the technical layer (midstream link), and the application layer (downstream link). Specifically, the basic layer includes the three pillars of computing power infrastructure, core hardware components, and the data resource system; the technical layer focuses on key technical modules such as algorithm research and development, platform construction, and tool development; the application layer concerns the development of smart terminal products and the implementation of industry scenario solutions. In identifying the categories of sample enterprises, this paper constructs a text classification algorithm model based on the QWEN-7B large language model (Bai et al., \u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e2023\u003c/span\u003e) and leverages natural language processing technology to extract semantic features and recognize patterns in the business scope text of the enterprise, ultimately achieving precise positioning and hierarchical attribution determination of the sample enterprises within the industrial chain.\u003c/p\u003e \u003cp\u003eBy analyzing data from sample enterprises, it is observed that the upstream sector primarily comprises high-emission companies dependent on traditional energy sources. Since basic hardware manufacturing and large-scale data center operations necessitate a continuous and stable supply of significant amounts of electricity, they predominantly rely on traditional energy sources, such as coal-fired power. The midstream sector can partially utilize renewable energy for technological research and development, as well as platform operations, but it still requires assurances from traditional energy sources for computing-intensive tasks. The downstream sector mainly consists of low-emission enterprises that utilize clean energy, since terminal applications have relatively low energy demands, and the companies' images and ESG requirements compel them to use clean energy to meet carbon neutrality objectives. Columns (3), (4), and (5) of Table\u0026nbsp;\u003cspan refid=\"Tab7\" class=\"InternalRef\"\u003e7\u003c/span\u003e display the estimated effects of the regional energy consumption structure on the upstream, midstream, and downstream segments of the industry. The findings reveal that the energy consumption structure significantly negatively impacts the downstream sector at the 1% level, while it positively affects the midstream sector at the 10% level. This indicates that the low-carbon city pilot policy fosters the growth of the regional artificial intelligence industry by optimizing the energy consumption structure. It mainly supports the development of the downstream portion of the industrial chain while hindering the growth of the midstream part. This may be due to significant technological substitution effects and the rigidities of factor substitution in the energy transformation process: downstream application scenarios can directly align with terminal market demands and quickly achieve clean energy substitution through flexible energy adaptations (such as deploying distributed computing power), thereby benefiting from policy-driven technological diffusion gains. In contrast, midstream technology-focused companies are restricted by the path dependency of high-precision algorithm research and development on a stable energy supply, facing dual pressures from equipment upgrades and rising R\u0026amp;D costs in the short term, which leads to the variability of policy transmission.\u003c/p\u003e \u003cp\u003e \u003cb\u003eMechanism Analysis: Green Technology Innovation.\u003c/b\u003e To test the potential mechanism behind the innovation compensation effect, this paper evaluates whether the low-carbon city pilot policy can stimulate the growth of the regional artificial intelligence industry by promoting green technology innovation. Green technology innovation may manifest as the low-carbon transformation of production processes, the upgrading of clean technology equipment, and the reconstruction of green product systems. The low-carbon city pilot policy encourages enterprises to move beyond traditional technology pathways and adopt green innovation by establishing a carbon quota trading mechanism and a tiered carbon tax system. This includes developing intelligent energy consumption management systems, deploying industrial Internet of Things (IIoT) energy efficiency optimization platforms, and implementing carbon footprint real-time monitoring technology. Considering the availability of urban green innovation technology data, this paper uses the logarithm of the number of green inventions filed by the city in that year as a proxy variable for green innovation technology. This indicator not only reflects the knowledge output density of technological innovation but also addresses the quality heterogeneity problem associated with utility model patents (Aghion et al., \u003cspan citationid=\"CR2\" class=\"CitationRef\"\u003e2016\u003c/span\u003e).\u003c/p\u003e \u003cp\u003eTable\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, column (1) presents the estimated results concerning the impact of the low-carbon city pilot policy on green technology innovation. The coefficient estimate for LCPP is 0.161, and it passes the significance test at the 5% level. This indicates that the low-carbon city pilot policy has significantly stimulated green technology innovation. Furthermore, this paper extends the analysis to investigate the impact of green technology innovation on the development of the regional artificial intelligence industry. Table\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, columns (2), (3), (4), and (5) illustrate that the promotional effects of green technology innovation on the entire industrial chain, the upstream basic layer, and the downstream application layer are 0.035 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.1), 0.021 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.1), and 0.047 (p\u0026thinsp;\u0026lt;\u0026thinsp;0.01), respectively. However, the effect on the midstream technology layer did not pass the significance test (β\u0026thinsp;=\u0026thinsp;0.009, p\u0026thinsp;\u0026gt;\u0026thinsp;0.1). This may be attributed to the \u0026lsquo;double-end driving effect\u0026rsquo; of technological innovation on the industrial chain: upstream basic layer enterprises have significantly reduced their marginal emission reduction costs (MAC) due to policy subsidies and tax incentives (such as subsidies for high-efficiency chip design), which directly support clean technology research and development. In contrast, downstream application layer enterprises can quickly integrate green technology into smart terminal products (such as new energy autonomous driving systems) through flexible resource reconfiguration, allowing them to gain market premiums via product differentiation. Conversely, the midstream technology layer faces limitations due to the path-dependence characteristics of algorithm models (Arthur, \u003cspan citationid=\"CR1\" class=\"CitationRef\"\u003e1989\u003c/span\u003e), making it difficult to break out of the existing technology framework and achieve green transformation in the short term, which hinders the transmission of innovation dividends.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab8\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 8\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eMechanism analysis: Green Technology Innovation\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(5)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003e\u003cb\u003eVARIABLES\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e\u003cb\u003eGTI\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e\u003cb\u003eDAII\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e\u003cb\u003eUpstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e\u003cb\u003eMidstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e\u003cb\u003eDownstream\u003c/b\u003e\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCCP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.161**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(2.49)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e\u0026nbsp;\u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eGTI\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.035***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.057***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.004\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.021*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(6.80)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3.17)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(0.80)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(1.67)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e4,543\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.260\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.884\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.601\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.114\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.136\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance. levels at 1%, 5% and 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eHeterogeneity analysis: Urban development level.\u003c/b\u003e Cities at various stages of development showcase differences in socioeconomic factors like population size, economic base, industrial structure, and capacity for policy implementation. Therefore, this study explores the diverse impacts of low-carbon city pilot policies across two dimensions: city tiers and location within the Yangtze River Economic Belt. Based on the \u0026lsquo;2022 City Commercial Attractiveness Rankings\u0026rsquo; by Yicai, we categorize the sample cities into three groups - first-tier, second-tier, and third-tier or below - for heterogeneity analysis. Additionally, given the strategic significance of the Yangtze River Economic Belt, we further divide the sample into groups for the Yangtze region and non-Yangtze region for supplementary testing.\u003c/p\u003e \u003cp\u003eColumns (1)-(3) of Table\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e present regression results for first-tier, second-tier, and third-tier cities, respectively. The findings indicate that low-carbon city pilot policies significantly promote the development of the AI industry in third-tier and below cities, whereas the effects on first- and second-tier cities are insignificant. Columns (4) and (5) of Table\u0026nbsp;\u003cspan refid=\"Tab9\" class=\"InternalRef\"\u003e9\u003c/span\u003e display results for cities within and outside the Yangtze Economic Belt, revealing significant positive effects for cities within the Belt but insignificant effects for those beyond it. This may be due to the fact that, compared to the established industrial systems in first- and second-tier cities, as well as in Yangtze Economic Belt cities, emerging industries like AI in lower-tier and non-Belt cities are relatively smaller. This smaller scale allows policy incentives to concentrate limited resources for rapid industrial clustering, thereby facilitating a quick breakthrough of scale economy thresholds and enabling rapid industrial development.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab9\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 9\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHeterogeneity analysis: Urban development level\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(5)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.202\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.403\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.034**\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.296\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.526*\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(0.24)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(1.12)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(2.27)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(1.47)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(1.86)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e374\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1639\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e2100\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e1544\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e2568\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.744\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.413\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.611\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.510\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.495\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance. levels at 1%, 5% and. 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eHeterogeneity analysis: Geographical location.\u003c/b\u003e Given China\u0026rsquo;s vast territory, cities show considerable regional differences in natural geographic elements such as topography, climate, and resource availability. Thus, this study investigates whether the impacts of low-carbon city pilot policies exhibit geographical diversity, categorizing the sample into eastern, central, and western regions.\u003c/p\u003e \u003cp\u003eAs shown in Table\u0026nbsp;\u003cspan refid=\"Tab10\" class=\"InternalRef\"\u003e10\u003c/span\u003e, Column (3), which represents cities in the western region, shows a regression coefficient of 0.689, significant at the 10% level. However, the coefficients for cities in the eastern region (Column 1) and central region (Column 2) are insignificant, indicating that the positive effects of low-carbon city pilot policies on AI industry development are present only in the western region. This empirically illustrates the geographical heterogeneity in the effects of policy implementation, likely closely related to the industrial foundation and resource endowment characteristics of the western region. Given the later start of AI industry development in western regions, the marginal effects of policy incentives and resource input may be more pronounced, reflecting differentiated policy implementation effects against the backdrop of China\u0026rsquo;s unbalanced regional development.\u003c/p\u003e \u003cp\u003e \u003cdiv class=\"gridtable\"\u003e\u003ctable float=\"Yes\" id=\"Tab10\" border=\"1\"\u003e \u003ccaption language=\"En\"\u003e \u003cdiv class=\"CaptionNumber\"\u003eTable 10\u003c/div\u003e \u003cdiv class=\"CaptionContent\"\u003e \u003cp\u003eHeterogeneity analysis: Geographical location \u0026amp; Industrial Dependence\u003c/p\u003e \u003c/div\u003e \u003c/caption\u003e \u003ccolgroup cols=\"6\"\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c1\" colnum=\"1\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c2\" colnum=\"2\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c3\" colnum=\"3\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c4\" colnum=\"4\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c5\" colnum=\"5\"\u003e\u003c/div\u003e \u003cdiv align=\"left\" class=\"colspec\" colname=\"c6\" colnum=\"6\"\u003e\u003c/div\u003e \u003cthead\u003e \u003ctr\u003e \u003cth align=\"left\" colname=\"c1\"\u003e \u003cp\u003eVARIABLES\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(2)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(3)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(4)\u003c/p\u003e \u003c/th\u003e \u003cth align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(5)\u003c/p\u003e \u003c/th\u003e \u003c/tr\u003e \u003c/thead\u003e \u003ctbody\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eLCPP\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.337\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.390\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.689*\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.411***\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.493\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e\u0026nbsp;\u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e(1.25)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e(1.44)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e(1.78)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e(2.22)\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e(1.51)\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eControls\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eObservations\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e1762\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e1162\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e1,189\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e2,577\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e1,356\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eR-squared\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003e0.571\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003e0.645\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003e0.406\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003e0.483\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003e0.628\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eid fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003ctr\u003e \u003ctd align=\"left\" colname=\"c1\"\u003e \u003cp\u003eyear fe\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c2\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c3\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c4\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c5\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003ctd align=\"left\" colname=\"c6\"\u003e \u003cp\u003eyes\u003c/p\u003e \u003c/td\u003e \u003c/tr\u003e \u003c/tbody\u003e \u003c/colgroup\u003e \u003ctfoot\u003e \u003ctr\u003e\u003ctd colspan=\"6\"\u003eNotes: The standard errors are indicated in parentheses. \u003csup\u003e⁎⁎⁎\u003c/sup\u003e, \u003csup\u003e⁎⁎\u003c/sup\u003e, and \u003csup\u003e⁎\u003c/sup\u003e denote significance. levels at 1%, 5% and. 10%, respectively.\u003c/td\u003e\u003c/tr\u003e \u003c/tfoot\u003e \u003c/table\u003e\u003c/div\u003e \u003c/p\u003e \u003cp\u003e \u003cb\u003eHeterogeneity analysis: Industrial Dependence.\u003c/b\u003e Resource-based cities and old industrial bases that are overly dependent on traditional industrial development models, while relatively slow in fostering emerging industries such as AI, confront structural imbalances that contribute to their \u0026lsquo;regional collapse\u0026rsquo; predicament. Following Zhang et al. (\u003cspan citationid=\"CR51\" class=\"CitationRef\"\u003e2024\u003c/span\u003e), we use the 2015 mean secondary industry proportion across all cities as a benchmark to categorize sample cities into \u0026lsquo;high secondary industry proportion\u0026rsquo; and \u0026lsquo;low secondary industry proportion\u0026rsquo; groups, as shown in Columns (4) and (5) of Table\u0026nbsp;\u003cspan refid=\"Tab8\" class=\"InternalRef\"\u003e8\u003c/span\u003e, to investigate whether low-carbon city construction has a more significant impact on industry-dependent cities.\u003c/p\u003e \u003cp\u003eThe regression results indicate that low-carbon city pilot policies significantly enhance AI industry development in cities with higher industrial dependence, yet show no significant effect on cities with lower industrial dependence. This might be due to the policies encouraging new-old momentum conversion in industry-dependent cities, enabling a reasonable transformation and optimization of existing industrial infrastructure, technological accumulation, and talent reserves for the growth of emerging industries, thus accelerating AI industry development.\u003c/p\u003e"},{"header":"Discussion","content":"\u003cp\u003eUsing panel data from 285 Chinese cities between 2007 and 2022, this study evaluated the impact of China\u0026rsquo;s LCPP program on the development of the urban AI industry through a staggered DID approach. The main results indicate that the implementation of China\u0026rsquo;s LCPP plan significantly positively influences the growth of the urban artificial intelligence industry. Additionally, the study conducted robustness tests using a placebo test, instrumental variable method, PSM-DID, replacing explained variables, altering clustering methods, adjusting the timing of LCPP occurrence, and controlling for other policy influences. The estimated results are consistent with the main findings, demonstrating the robustness of our conclusions.\u003c/p\u003e \u003cp\u003eAdditionally, we found that environmental regulation may drive the development of the artificial intelligence industry through two mechanisms: the first is the reconstruction of the energy consumption structure. When the price of traditional energy sources such as coal-fired power experiences a rigid increase due to policy tools like environmental taxes and quota restrictions, the large-scale utilization of green energy reshapes the electricity cost curve with its low marginal cost. This creates a price depression for low-carbon energy, attracting the agglomeration and growth of the artificial intelligence industry through low-cost electricity, predominantly driving the clustering of the downstream segments of the low-emission industrial chain. The second is the green technology innovation mechanism. Green technology innovation, prompted by environmental regulation, has notable positive externality characteristics. Its knowledge spillover effect can lower the threshold for technology adoption in related industries, thereby assisting the artificial intelligence industry in acquiring and implementing energy-saving technologies at a reduced cost, which, in turn, promotes its growth, particularly enhancing the development of the industry\u0026rsquo;s high-emission upstream segments and low-emission downstream segments. Finally, through heterogeneity tests, we found that the positive impact of the LCPP plan on the urban artificial intelligence industry is only significant in the western region, in third-tier cities and below, as well as in cities outside the Yangtze River Economic Belt. Moreover, the positive impact of the LCPP plan on the development of the artificial intelligence industry in heavily industrialized cities is more pronounced. This finding indicates that the role of the LCPP plan in fostering the development of the urban artificial intelligence industry is influenced by urban characteristics and does not apply uniformly across all cities.\u003c/p\u003e \u003cp\u003e \u003cb\u003eTheoretical implications.\u003c/b\u003eThe theoretical implications of this study are as follows:\u003c/p\u003e \u003cp\u003eFirst, existing studies primarily concentrate on the effects of environmental regulation policies on the transformation and upgrading of traditional industries, such as promoting technological innovation in enterprises by restricting high pollution emissions (Yang, Jahanger, and Hossain, \u003cspan citationid=\"CR47\" class=\"CitationRef\"\u003e2023\u003c/span\u003e), or examining the influence of environmental regulation on enterprise productivity and competitiveness (Dagestani et al., \u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e2023\u003c/span\u003e). This study broadens the research perspective to include strategic emerging industries and empirically tests the role of environmental regulation in advancing the development of the artificial intelligence sector, which not only addresses a gap in existing research but also enhances the theoretical understanding of the economic impacts of environmental regulation policies.\u003c/p\u003e \u003cp\u003eSecond, current research on the development of the artificial intelligence industry primarily focuses on technological advancements, talent aggregation, and capital investment (Rashid \u0026amp; Kausik, \u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e2024\u003c/span\u003e; Zhang, \u003cspan citationid=\"CR52\" class=\"CitationRef\"\u003e2023\u003c/span\u003e; Babina, T et al., \u003cspan citationid=\"CR5\" class=\"CitationRef\"\u003e2024\u003c/span\u003e). However, the role of policy factors in industrial development cannot be overlooked. This study introduces an innovative perspective on environmental regulation policies. By examining the implementation effects and mechanisms of LCPP, it highlights the significant role of policy factors in promoting the development of the artificial intelligence industry. This suggests that when analyzing the driving forces behind the growth of the artificial intelligence sector, we should consider not only technical and market factors but also the influence of the policy environment and institutional design.\u003c/p\u003e \u003cp\u003eThird, this study reveals the diverse impact of low-carbon policies on the development of urban artificial intelligence industries, providing empirical evidence for the implementation of targeted and effective policies. The study found that the positive effects of low-carbon policies on the artificial intelligence sector are more pronounced in the western regions, third-tier and smaller cities, areas outside the Yangtze River Economic Belt, and heavily industrialized cities. This finding illustrates that when implementing low-carbon policies to promote the growth of emerging industries, it is crucial to fully consider regional differences and urban characteristics and to adopt strategies tailored to local conditions.\u003c/p\u003e \u003cp\u003e \u003cb\u003ePolicy implications.\u003c/b\u003e The findings of this study offer important insights for formulating low-carbon policies and promoting AI industry development. Specific policy implications include:\u003c/p\u003e \u003cp\u003eFirst, governments should utilize low-carbon city pilot programs as an essential tool for advancing AI industry development. Drawing from China\u0026rsquo;s successful experiences, we recommend that governments: (1) Establish national-level low-carbon city pilot projects, select pilot cities through competitive evaluation, and provide appropriate policy support and financial incentives; (2) Create a policy framework that encourages the synergistic development of low-carbon initiatives and the AI industry, facilitating the innovative application of AI technologies in low-carbon areas such as energy management and environmental monitoring; (3) Develop a comprehensive evaluation system for low-carbon cities, incorporating the level of AI industry development into the evaluation criteria, and guiding cities to prioritize the cultivation of emerging industries during their low-carbon transition.\u003c/p\u003e \u003cp\u003eSecond, the regional heterogeneity analysis reveals significant variations in the effectiveness of low-carbon pilot policies across different regions. In China, when promoting low-carbon city construction, the development needs of less-developed regions and industrial transformation cities should take precedence. Specifically: (1) Develop differentiated support policies for low-carbon pilots, offering preferential funding and project layout assistance to less-developed regions; (2) Create specialized support plans for industrial transformation in traditional industrial cities, encouraging the deep integration of traditional industries with AI technologies through the establishment of industrial transformation funds; (3) Establish regional cooperation mechanisms to enable experience sharing and technical collaboration on low-carbon technologies and the AI industry between developed and less-developed regions.\u003c/p\u003e \u003cp\u003eThird, based on the analysis of the policy impact mechanism, it is recommended to enhance the role of low-carbon policies in promoting the artificial intelligence industry from multiple perspectives: (1) It is suggested to deepen the market-oriented reform of energy prices, establish a composite pricing system, expand the peak-valley electricity price difference, and strengthen the guiding influence of price signals; (2) It is proposed to create a technology collaborative innovation consortium, establish a \u0026lsquo;dual carbon-AI\u0026rsquo; national laboratory, overcome common technological barriers, and implement a rapid review system for green patents and compulsory licensing; (3) Implement a policy promotion strategy that varies in time and space, and develop corresponding policies based on the characteristics of different regions, such as focusing on smart grids in eastern coastal cities, laying out green computing infrastructure in central and western energy hubs, and promoting digitalization and carbon reduction in traditional manufacturing within the old industrial base in Northeast China.\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e\u003cstrong\u003eAuthor contributions\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eauthor one: Data curation, Formal analysis, Investigation, Visualization, Writing\u0026mdash;original\u003c/p\u003e\n\u003cp\u003edraft, Writing\u0026mdash;review \u0026amp; editing. author two: Formulate research questions, Data curation, Writing\u0026mdash;review \u0026amp; editing. author three.: Supervision, Writing\u0026mdash;review \u0026amp; editing.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eData availability\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe datasets generated and analyzed during the current study are available from the corresponding author upon reasonable request.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eEthical approval\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe study is exempt from review by the Ethics Committee since it is classified as\u003c/p\u003e\n\u003cp\u003enonhuman subject research, thereby waiving the need for informed consent.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eInformed consent\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThis article does not contain studies involving human participants engaged by any of the\u003c/p\u003e\n\u003cp\u003eauthors.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAdditional information\u0026nbsp;\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eCorrespondence and requests for materials should be addressed to author one.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eArthur, W. B. (1989). Competing technologies, increasing returns, and lock-in by historical events. The economic journal, 99(394), 116-131.\u003c/li\u003e\n\u003cli\u003eAghion, P., Dechezlepr\u0026ecirc;tre, A., Hemous, D., Martin, R., \u0026amp; Van Reenen, J. (2016). Carbon taxes, path dependency, and directed technical change: Evidence from the auto industry. Journal of Political Economy, 124(1), 1-51.\u003c/li\u003e\n\u003cli\u003eAntoniou, F., \u0026amp; Mageiropoulos, T. (2024). Ranking the Barriers to the Energy Upgrading of Buildings Using the Best-Worst Method. Sustainability, 16(22), 10143.\u003c/li\u003e\n\u003cli\u003eBai, J., Bai, S., Chu, Y., Cui, Z., Dang, K., Deng, X., ... \u0026amp; Zhu, T. (2023). Qwen technical report. arxiv preprint arxiv:2309.16609.\u003c/li\u003e\n\u003cli\u003eBabina, T., Fedyk, A., He, A., \u0026amp; Hodson, J. (2024). Artificial intelligence, firm growth, and product innovation. Journal of Financial Economics, 151(1), 103745.\u003c/li\u003e\n\u003cli\u003eBeck, T., Levine, R., \u0026amp; Levkov, A. (2010). Big bad banks? The winners and losers from bank deregulation in the United States. The Journal of Finance, 65(5), 1637\u0026ndash;1667.\u003c/li\u003e\n\u003cli\u003eBertrand, M., \u0026amp; Mullainathan, S. (2003). Enjoying the quiet life? Corporate governance and managerial preferences. Journal of Political Economy, 111(5), 1043\u0026ndash;1075.\u003c/li\u003e\n\u003cli\u003eChen, H., Feng, L., \u0026amp; Sun, X. (2024). Beyond central-local relations: the introduction of a new perspective on China\u0026rsquo;s environmental governance model. Humanities and Social Sciences Communications, 11(1), 1-12.\u003c/li\u003e\n\u003cli\u003eChen, M. (2022). A study of low-carbon development, urban innovation and industrial structure upgrading in China. International Journal of Low-Carbon Technologies, 17(2), 185\u0026ndash;195.\u003c/li\u003e\n\u003cli\u003eCyranoski, D. (2018). Chinese firms enter the battle for AI talent. Nature, 553(7688), 260\u0026ndash;261.\u003c/li\u003e\n\u003cli\u003eChetty, R., Looney, A., \u0026amp; Kroft, K. (2009). Salience and taxation: Theory and evidence. American Economic Review, 99(4), 1145\u0026ndash;1177.\u003c/li\u003e\n\u003cli\u003eChong, Z., Qin, C., \u0026amp; Ye, X. (2017). Environmental regulation and industrial structure change in China: Integrating spatial and social network analysis. Sustainability, 9(8), 1465.\u003c/li\u003e\n\u003cli\u003eDong, X., Song, S., \u0026amp; Zhu, H. (2011). Industrial structure and economic fluctuation: Evidence from China. The Social Science Journal, 48(3), 468\u0026ndash;477.\u003c/li\u003e\n\u003cli\u003eDagestani, A. A., Shang, Y., Schneider, N., Cifuentes-Faura, J., \u0026amp; Zhao, X. (2023). Porter in China: A quasi-experimental view of market-based environmental regulation effects on firm performance. Energy Economics, 126(10), 106966.\u003c/li\u003e\n\u003cli\u003eDong, Y., Tian, J., \u0026amp; Wen, Q. (2022). Environmental regulation and outward foreign direct investment: Evidence from China. China Economic Review, 76(4), 101877.\u003c/li\u003e\n\u003cli\u003eDong, Z. Q., \u0026amp; Wang, H. (2019). The \u0026quot;local-neighborhood\u0026quot; green technology progress effect of environmental regulation. China Industrial Economy, (1), 100-118. \u003c/li\u003e\n\u003cli\u003eFan, F., \u0026amp; Zhang, X. (2021). Transformation effect of resource-based cities based on PSM-DID model: An empirical analysis from China. Environmental Impact Assessment Review, 91, 106648.\u003c/li\u003e\n\u003cli\u003eFahad, S., Bai, D., Liu, L., \u0026amp; Baloch, Z. A. (2022). Heterogeneous impacts of environmental regulation on foreign direct investment: Do environmental regulation affect FDI decisions? Environmental Science and Pollution Research, 29(4), 5092\u0026ndash;5104.\u003c/li\u003e\n\u003cli\u003eGao, Y. (2022). Unleashing the mechanism among environmental regulation, artificial intelligence, and global value chain leaps: A roadmap toward digital revolution and environmental sustainability. Environmental Science and Pollution Research, 30(10), 28107\u0026ndash;28117.\u003c/li\u003e\n\u003cli\u003eHong Kong Productivity Council and Hong Kong Institute of Economics and Business Strategy at HKU Business School. 2024. \u0026quot;Hong Kong AI Industry Development Study.\u0026quot; Hong Kong: HKPC and HIEBS. https://www.hkpc.org/sites/default/files/2024-04/hkpc_hku_ai_industry_development_study_en.pdf. Accessed November 20\u003c/li\u003e\n\u003cli\u003eHaas, A., \u0026amp; Osland, L. (2014). Commuting, migration, housing and labour markets: Complex interactions. Urban Studies, 51(3), 463\u0026ndash;476.\u003c/li\u003e\n\u003cli\u003eHong, M., Chen, S., \u0026amp; Zhang, K. (2021). Impact of the \u0026quot;Low-Carbon City Pilot\u0026quot; policy on energy intensity based on the empirical evidence of Chinese cities. Frontiers in Environmental Science, 9(7), Article 717737.\u003c/li\u003e\n\u003cli\u003eHu, Y., Dai, X., \u0026amp; Zhao, L. (2022). Digital finance, environmental regulation, and green technology innovation: An Empirical Study of 278 Cities in China. Sustainability, 14(14), 8652.\u003c/li\u003e\n\u003cli\u003eInternational Energy Agency. 2023. \u0026quot;Emissions Grew in 2023, but Clean Energy Is Limiting the Growth - CO2 Emissions in 2023 - Analysis - IEA.\u0026quot; Accessed November 20. https://www.iea.org/reports/co2-emissions-in-2023/emissions-grew-in-2023-but-clean-energy-is-limiting-the-growth\u003c/li\u003e\n\u003cli\u003eKrugman, P. (1992). Geography and trade. MIT press.\u003c/li\u003e\n\u003cli\u003eLi, C., Liang, F., Liang, Y., \u0026amp; Wang, Z. (2023). Low-carbon strategy, entrepreneurial activity, and industrial structure change: evidence from a quasi-natural experiment. Journal of Cleaner Production, 427, 139183.\u003c/li\u003e\n\u003cli\u003eLi, S., Zheng, X., Liao, J., \u0026amp; Niu, J. (2024). Low-carbon city pilot policy and corporate environmental performance: Evidence from a quasi-natural experiment. International Review of Economics \u0026amp; Finance, 89, 1248-1266.\u003c/li\u003e\n\u003cli\u003eLiu, Y., Peng, Y., Wang, W., Liu, S., \u0026amp; Yin, Q. (2024). Does the pilot zone for green finance reform and innovation policy improve urban green total factor productivity? The role of digitization and technological innovation. Journal of Cleaner Production, 471, 143365.\u003c/li\u003e\n\u003cli\u003eLuqman, M., Rayner, P. J., \u0026amp; Gurney, K. R. (2023). On the impact of urbanisation on CO2 emissions. NPJ Urban Sustainability, 3(1), 6.\u003c/li\u003e\n\u003cli\u003eMi, T., \u0026amp; Li, T. (2024). Industrial Intelligence and Carbon Emission Reduction: Evidence from China\u0026apos;s Manufacturing Industry. Sustainability (2071-1050), 16(15).\u003c/li\u003e\n\u003cli\u003eMakiela, K., \u0026amp; Ouattara, B. (2018). Foreign direct investment and economic growth: Exploring the transmission channels. Economic Modelling, 72, 296-305.\u003c/li\u003e\n\u003cli\u003ePei, Y., Zhu, Y., Liu, S., Wang, X., \u0026amp; Cao, J. (2019). Environmental regulation and carbon emission: The mediation effect of technical efficiency. Journal of Cleaner Production, 236, 117599.\u003c/li\u003e\n\u003cli\u003ePopp, D. (2006). International innovation and diffusion of air pollution control technologies: the effects of NOX and SO2 regulation in the US, Japan, and Germany. Journal of Environmental Economics and Management, 51(1), 46-71.\u003c/li\u003e\n\u003cli\u003ePorter, M. E., \u0026amp; Linde, C. V. D. (1995). Toward a new conception of the environment-competitiveness relationship. Journal of economic perspectives, 9(4), 97-118.\u003c/li\u003e\n\u003cli\u003eRashid, A. B., \u0026amp; Kausik, A. K. (2024). AI revolutionizing industries worldwide: A comprehensive overview of its diverse applications. Hybrid Advances, 100277.\u003c/li\u003e\n\u003cli\u003eRen, X., Zhang, X., Yan, C., \u0026amp; Gozgor, G. (2022). Climate policy uncertainty and firm-level total factor productivity: Evidence from China. Energy Economics, 113, 106209.\u003c/li\u003e\n\u003cli\u003eRen, H., Gu, G., \u0026amp; Zhou, H. (2022). Assessing the low-carbon city pilot policy on carbon emission from consumption and production in China: how underlying mechanism and spatial spillover effect?. Environmental Science and Pollution Research, 29(47), 71958-71977.\u003c/li\u003e\n\u003cli\u003eSi, X. (2024). A study on the impact of talent introduction policies on urban mobility: A literature review. Journal of Education, Humanities and Social Sciences, 39(November), 154\u0026ndash;160.\u003c/li\u003e\n\u003cli\u003eSun, X., Song, Y., \u0026amp; Zhao, P. Y. (2022). The impact of artificial intelligence on the employment of heterogeneous labor force: From the perspective of labor supply. Economic Problem Exploration, (02), 171\u0026ndash;190.\u003c/li\u003e\n\u003cli\u003eTao, Weiliang, Shimei Weng, Xueli Chen, Fawaz Baddar ALHussan, and Malin Song, \u0026ldquo;Artificial Intelligence-Driven Transformations in Low-Carbon Energy Structure: Evidence from China.\u0026rdquo; Energy Economics, 2024, 136 (August), 107719.\u003c/li\u003e\n\u003cli\u003eWu, L., Sun, L., Qi, P., Ren, X., \u0026amp; Sun, X. (2021). Energy endowment, industrial structure upgrading, and CO2 emissions in China: Revisiting resource curse in the context of carbon emissions. Resources Policy, 74, 102329.\u003c/li\u003e\n\u003cli\u003eWu, G., Sun, M., \u0026amp; Feng, Y. (2024). How does the new environmental protection law affect the environmental social responsibility of enterprises in Chinese heavily polluting industries?. Humanities and Social Sciences Communications, 11(1), 1-14.\u003c/li\u003e\n\u003cli\u003eWang, X., Zhou, J., \u0026amp; Maitlo, Q. (2022). Effect of green technology innovation on the upgrading of the manufacturing value chain: evidence from China. Frontiers in Environmental Science, 10, 994323.\u003c/li\u003e\n\u003cli\u003eXing, Z., Guo, J., Zhang, Z., Xue, T., Yang, M., \u0026amp; Wu, W. (2024). Research on the Impact of Environmental Inequality on Labor Mobility\u0026mdash;A Study Based on the China General Social Survey (CGSS). Sustainability (2071-1050), 16(22).\u003c/li\u003e\n\u003cli\u003eYang, S., Jahanger, A., \u0026amp; Hossain, M. R. (2023). How effective has the low-carbon city pilot policy been as an environmental intervention in curbing pollution? Evidence from Chinese industrial enterprises. Energy Economics, 118, 106523.\u003c/li\u003e\n\u003cli\u003eYin, H., \u0026amp; Su, W. (2024). Industrial synergy agglomeration, urban innovation capacity, and advanced manufacturing development. Economies, 12(5), 117.\u003c/li\u003e\n\u003cli\u003eYu, X., \u0026amp; Wang, P. (2021). Economic effects analysis of environmental regulation policy in the process of industrial structure upgrading: Evidence from Chinese provincial panel data. Science of the Total Environment, 753, 142004.\u003c/li\u003e\n\u003cli\u003eZhang, L., Li, Y., Kung, C. C., Wu, B., \u0026amp; Zhang, C. (2023). Impact of new talent settlement policy on housing prices: Evidence from 70 large and medium-sized Chinese cities. PloS one, 18(3), e0280317.\u003c/li\u003e\n\u003cli\u003eZhang, L., Lin, G., Lyu, X., \u0026amp; Su, W. (2024). Suppression or promotion: research on the impact of industrial structure upgrading on urban economic resilience. Humanities and Social Sciences Communications, 11(1), 1-14.\u003c/li\u003e\n\u003cli\u003eZhang, Z. (2023). The impact of the artificial intelligence industry on the number and structure of employments in the digital economy environment. Technological Forecasting and Social Change, 197, 122881.\u003c/li\u003e\n\u003cli\u003eZhu, C., \u0026amp; Lee, C. C. (2022). The effects of low-carbon pilot policy on technological innovation: Evidence from prefecture-level data in China. Technological Forecasting and Social Change, 183, 121955.\u003c/li\u003e\n\u003cli\u003eZou, C., Huang, Y., Wu, S., \u0026amp; Hu, S. (2022). Does \u0026ldquo;low-carbon city\u0026rdquo; accelerate urban innovation? Evidence from China. Sustainable Cities and Society, 83, 103954.\u003c/li\u003e\n\u003cli\u003eZhou, Q., \u0026amp; Qi, Z. (2023). Urban economic resilience and human capital: An exploration of heterogeneity and mechanism in the context of spatial population mobility. Sustainable Cities and Society, 99, 104983.\u003c/li\u003e\n\u003cli\u003eZhou, Y. H., Yang, L., \u0026amp; Jiang, S. S. (2023). Binding carbon emission reduction and employment: An investigation based on changes in enterprise and regional labor force. Economic Research, 58(7), 104\u0026ndash;120.\u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":false,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-6253945/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-6253945/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"The ‘low-carbon city’ pilot policy is a crucial measure for China to advance green and low-carbon development. The rapid growth of the artificial intelligence industry may lead to significant energy consumption and carbon emissions, prompting discussions about the potential conflict between low-carbon city construction and the growth of the artificial intelligence sector. This study utilizes China’s ‘low-carbon city’ pilot policy as a natural experiment, employs a staggered double difference model (Staggered DID), empirically tests the impact of the policy on the development of the urban artificial intelligence industry, and thoroughly explores its mechanisms and heterogeneity characteristics. The findings reveal that the low-carbon city pilot policy has not hindered the growth of the urban artificial intelligence industry. Compared to non-pilot cities, the policy's implementation has resulted in an average increase of approximately 29.4% in the size of the artificial intelligence industry in pilot cities. By examining the energy consumption characteristics across different segments of the artificial intelligence industry chain, this study identifies that the policy has a varied effect in promoting the development of the regional artificial intelligence industry by altering the urban energy consumption structure, primarily driving the clustering of the downstream segments of the low-emission industry chain. Additionally, enhancing the level of urban green innovation mainly supports the development of high-emission upstream segments and low-emission downstream segments of the industry. Heterogeneity analysis further indicates that the policy's positive effects significantly differ across various regions, city tiers, economic zones, and cities with differing levels of industrialization. This study not only confirms the compatibility of low-carbon city construction with the growth of the artificial intelligence industry but also offers vital policy insights for balancing economic growth with carbon emission targets.","manuscriptTitle":"Are low-carbon cities and the development of the artificial intelligence industry contradictory?Evidence from China","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2025-04-25 10:42:51","doi":"10.21203/rs.3.rs-6253945/v1","editorialEvents":[{"type":"communityComments","content":0},{"type":"decision","content":"Revision requested","date":"2025-06-25T13:49:02+00:00","index":"","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-05-28T06:23:13+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"265798724904223810014874522876147761573","date":"2025-05-27T04:21:38+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"326652642593506942702459504441534420351","date":"2025-05-26T22:58:36+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"101175132749874353008222082539977240948","date":"2025-04-18T08:45:28+00:00","index":"hide","fulltext":""},{"type":"editorInvitedReview","content":"","date":"2025-04-18T07:29:56+00:00","index":"hide","fulltext":""},{"type":"reviewerAgreed","content":"325901407407521748532688206887544870210","date":"2025-04-18T05:32:33+00:00","index":"hide","fulltext":""},{"type":"reviewersInvited","content":"","date":"2025-04-18T05:26:59+00:00","index":"","fulltext":""},{"type":"editorInvited","content":"","date":"2025-04-17T12:01:39+00:00","index":"","fulltext":""},{"type":"editorAssigned","content":"","date":"2025-04-17T11:53:01+00:00","index":"","fulltext":""},{"type":"checksComplete","content":"","date":"2025-03-31T22:22:32+00:00","index":"","fulltext":""},{"type":"submitted","content":"Humanities and Social Sciences Communications","date":"2025-03-18T13:58:40+00:00","index":"","fulltext":""}],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"humanities-and-social-sciences-communications","isNatureJournal":false,"hasQc":true,"allowDirectSubmit":false,"externalIdentity":"palcomms","sideBox":"Learn more about [Humanities \u0026 Social Sciences Communications](http://www.nature.com/palcomms/)","snPcode":"41599","submissionUrl":"https://submission.springernature.com/new-submission/41599/3","title":"Humanities and Social Sciences Communications","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"stoa","reportingPortfolio":"Nature AJ","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"88a41978-c7d5-457b-92b2-fc6f7ef696df","owner":[],"postedDate":"April 25th, 2025","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":47354592,"name":"Social science/Economics"},{"id":47354593,"name":"Social science/Environmental studies"},{"id":47354594,"name":"Social science/Science technology and society"}],"tags":[],"updatedAt":"2026-04-16T07:38:27+00:00","versionOfRecord":[],"versionCreatedAt":"2025-04-25 10:42:51","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-6253945","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-6253945","identity":"rs-6253945","version":["v1"]},"buildId":"8U1c8b4HqxoKbykW_rLl7","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}