Broadband network is an important part of modern digital information development. By providing fast and reliable Internet access services, it provides an important support for high-quality development of modern economy. To accelerate the widespread adoption of broadband infrastructure, the Chinese government issued the Broadband China Strategy and Implementation Plan in August 2013, which clearly promoted the popularization of high-speed, safe and green digital infrastructure in China through regional pilot. Subsequently, three groups of pilot cities were identified from 2014 to 2016, gradually expanding the pilot scope of Internet broadband access business opening. The Broadband China strategy focuses on the construction of digital infrastructure, and meets the needs of the majority of economic entities for high-quality Internet services by improving network coverage and transmission speed. By adopting a number of policy measures, such as increasing capital investment, strengthening government guidance and support, promoting cooperation among regions and sectors, and strengthening talent training and technological innovation, Broadband China is able to promote the convergence and development of broadband networks and other infrastructures. Thus, the Broadband China strategy is an important measure to promote the development of China’s digital economy, and can be used as a quasi-natural experiment to assess the construction of China’s digital infrastructure.

The relationship between digital infrastructure and carbon reduction is essentially the relationship between the environmental effects of the digital economy. The digital economy is a new economic form formed by the information technology revolution and network infrastructure construction. Firstly, with the popularization of digital infrastructure centered on broadband network, it largely reduces the cost of spreading and searching for green and low-carbon information and increases the government’s ability to publicize the concept of urban green and low-carbon, which not only improves the city government’s ability to govern the environment, but also effectively reduces the city’s carbon emissions (Shahbaz et al., 2022). Secondly, with the construction of data infrastructure such as broadband China and cell phone base stations, digital media is increasingly entering residents’ lives. Through the promotion of digital media, residents’ attention to environmental issues is enhanced, guiding them to form green and low-carbon consumption concepts and lifestyles, promoting the implementation of green and low-carbon environmental protection concepts, and reducing urban carbon emissions (Strielkowski et al., 2022; Ma et al., 2023). Thirdly, digital infrastructure construction induces urban industrial agglomeration, which in turn promotes urban scale economy effect and technology diffusion effect, reduces information asymmetry between supply and demand sides of the market through the rapid dissemination of information, promotes the upgrading of industrial structure, and then promotes the occurrence of urban carbon emission reduction (Chen et al., 2025; Zhang et al., 2025).

Recent research has demonstrated that the environmental impacts of digital infrastructure development are becoming increasingly significant, with these effects being validated across multiple domains. Firstly, while rural areas historically experienced less digitalization influence compared to urban centers, recent digital infrastructure construction has driven substantial rural digital transformation. This shift has altered rural residents’ personal consumption patterns, including household energy consumption, which in turn promotes the use of clean energy in rural areas (Nguyen and Nasir, 2021). These findings have been confirmed by Yan et al. (2023). Secondly, in the industrial sector, where energy consumption such as electricity consumption and intensity is relatively high, digital infrastructure development has promoted technological optimization and industrial upgrading. These advancements have led to a marked reduction in industrial electricity demand, partially alleviating energy pressures while enhancing environmental efficiency (Williams, 2021; Wang et al., 2022). Thirdly, it promotes the green transformation of enterprises. Digital infrastructure helps enterprises optimize resource utilization efficiency through real-time data collection and analysis. On the other hand, it can integrate environmental monitoring data and provide accurate green management tools for enterprises, thereby promoting internal green upgrading (Zhong et al., 2022; Wu et al., 2023; Qian and Xi, 2024). In addition, the construction of digital infrastructure also has a significant impact on residents’ health (Xu et al., 2025).

The Broadband China strategy is an important element of digital infrastructure. A large number of empirical studies have been conducted on the relationship between the Strategy and environmental effects. Zou and Pan (2023) used broadband China as a quasi-natural experiment and empirically verified through DID and other methods that digital infrastructure construction can effectively reduce urban environmental pollution, and this pollution reduction effect is more pronounced in non resource-based cities, first and second tier cities, and eastern cities in China. Hong et al. (2023) build a time-varying difference-in-differences (time-varying DID) model based on the Broadband China pilot and find that digital infrastructure construction can significatively reduce prefecture-level energy intensity, and this environmental effect is more significant in Northeast China. For micro households, the broadband China strategy still has a significant impact. Zuo et al. (2024) through panel data on Chinese CFPS micro households, empirically tested based on the Broadband China strategy and found that digital infrastructure has the potential to indirectly increase household energy consumption through consumption upgrading and Internet development mechanisms, and that households in high-income, low-skilled, and rural areas, as well as those in areas with higher levels of development and marketization, are more deeply affected by digitalization, and household energy consumption will be more serious.

In summary, existing literature provide a good reference for this paper. Through the above literature, we have the following conclusions. Firstly, in terms of the choice of explanatory variables, a large number of studies have examined the economic impact or environmental impact brought about by the construction of digital infrastructure. Carbon emission reduction is an “extensive variable,” while carbon emission intensity, which is the amount of carbon emissions per unit of output value, is an “intensive variable.” Most studies tend to focus on one aspect, and there has not been sufficient attention given to measuring both as dependent variables simultaneously. Secondly, in the measurement of digital infrastructure construction, there are relatively few studies on the use of exogenous policy shocks. Research tends to favor selecting a single indicator or constructing a system of indicators, which may lead to certain biases in the results. Thirdly, most existing studies focus on the provincial or regional levels in China, or even at the enterprise level. Although some research involves the city level, the number of such studies is relatively limited. Fourthly, in heterogeneity studies, many researchers choose regional or city classifications, but these studies are relatively simplistic. There is a scarcity of research conducting heterogeneity analysis from the perspective of environmental regulation intensity. With the acceleration of urbanization in China, as an important economic carrier, cities have increasingly high requirements for environmental quality. Balancing the relationship between economic development and environmental quality is crucial for the sustainable development of cities. At the same time, the digital economy has become a new engine of economic development. It is worth exploring whether the construction of digital infrastructure, as an important component of the digital economy, can enable sustainable urban development.

Although the existing literature has provided valuable insights into the relationship between digital infrastructure and carbon emissions, there is still a need for further research to explore the specific mechanisms and heterogeneity of this relationship. This study aims to fill this gap by examining the carbon emission reduction effects of digital infrastructure construction in China, using the Broadband China strategy as a quasi-natural experiment.

Specifically, the development of digital infrastructure can enhance the level of urban green technology innovation in the following ways.

Firstly, the digital infrastructure construction of the Broadband China strategy has broken down information barriers, promoted communication and cooperation among governments, enterprises, and research institutes, facilitated spatial agglomeration of human capital, and provided human support for green technology innovation (Song et al., 2024). Secondly, the construction of digital infrastructure has laid the foundation for the construction of digital technology, accelerated the speed of information dissemination of the Internet and big data, reduced network search costs and information transmission costs, promoted the increase of returns to scale through information sharing and knowledge spillover, and provided the possibility for the development of urban green technology innovation (Cassia et al., 2020; Du and Wang, 2024). Thirdly, the construction of digital infrastructure has promoted the spatial architecture and application scope of big data, cloud computing and the Internet, and has derived new businesses and industries with digital technology as the core. This drives high-energy-consuming industries to achieve green transformation, optimizes energy consumption structures, supports the research and development of clean, low-carbon, and green technologies, and promotes green technology innovation (Veile et al., 2022).

According to Zhang et al. (2023), green technology innovation plays a significant role in accelerating urban carbon emission reduction. Firstly, green technology innovation facilitates the transformation of high-pollution industries, reduces industrial energy consumption, and decreases carbon emissions from industries (Wang, 2022). Secondly, green technology innovation promotes the innovation of production equipment and technology in enterprises, reduces production costs, lowers resource utilization and consumption, improves product green quality, thereby enhancing environmental quality and reducing enterprise carbon emissions (Ning et al., 2022). Thirdly, green technology innovation drives the energy revolution and accelerates the development of clean energy, supporting the transformation and upgrading of the energy consumption structure (Guo et al., 2021), which is conducive to building a clean and low-carbon modern energy system. The improvement of green technology innovation levels can promote the green and low-carbon transformation of energy, increase the proportion of clean energy consumption, and advance the construction of new energy systems, thereby continuously reducing energy consumption and urban carbon emissions (Madaleno et al., 2022; Wang and Lai, 2024).

Hypothesis 1

Digital infrastructure development reduces carbon emissions by improving the level of green technology innovation.

Firstly, the development of digital infrastructure has promoted industrial digitization. Through the “crowding-out effect,” it facilitates the transformation and upgrading of industries characterized by high energy consumption, high emissions, and high pollution. By improving the overall energy utilization structure of these sectors, it further reduces carbon emissions (Du et al., 2022; Gao X. et al., 2023). Secondly, digital infrastructure has a “penetration effect.” Relying on the strategy of Broadband China, promote digital technologies such as the Internet, big data, cloud computing, etc., promote the digital transformation of enterprises, improve the whole industry chain supervision system of enterprises’ supply from the source, middle end production, and terminal utilization, which is conducive to improving energy utilization efficiency, reducing resource waste, and achieving the effect of carbon emission reduction (Wang et al., 2024). Thirdly, the construction of digital infrastructure enhances the supply efficiency of digital elements and the integration efficiency of various resources. This fosters the emergence of new business models and industries, promoting the optimization of urban industrial structures. The integration of the digital economy with traditional industries accelerates urban digital industrialization and industrial digitization, drives the digital transformation of traditional sectors, and ultimately achieves urban industrial upgrading (Su et al., 2021; Das, 2024).

Upgrading the industrial structure involves adopting more scientific production methods and optimizing the production processes, thereby reducing unnecessary energy consumption and effectively mitigating carbon emissions (Zhao et al., 2023). Firstly, with the upgrading of industrial structure, it will particularly trigger technological upgrading within the manufacturing industry, drive productivity improvement, trigger the “endogenous innovation” effect, and achieve the improvement of urban energy efficiency and carbon reduction effect in China (Xi and Zhai, 2023). Secondly, upgrading the industrial structure will lead to the elimination of a group of high polluting and high emission enterprises, while low-carbon and environmentally friendly enterprises will develop and grow. The transformation of clean and pollution-free production models has reduced the emissions of pollutants from enterprises and achieved sustainable environmental development (You and Zhang, 2022; Gao Y. et al., 2023). Thirdly, upgrading the industrial structure will improve the efficiency of resource allocation in various departments, enhance the interaction between departments and industries, reduce the proportion of agriculture and industry, promote the development of the tertiary industry, and effectively curb urban carbon emissions through “economic servitization” (Xue et al., 2022; Fu et al., 2024).

Hypothesis 2

Digital infrastructure construction reduces carbon emissions by upgrading industrial structure. The Conceptual framework see Figure 1.

The selection of Broadband China pilot cities is mostly affected by broadband access capability and penetration rate, which are not logically affected by the level of prefecture-level carbon emissions and carbon intensity. It provides a good opportunity to exclude the reverse causality problem. Difference-of-differences models are often used to assess the effect of policy implementation, and the traditional DID is only suitable for assessing the policy effect of a single time point shock. However, the approval of the Broadband China demonstration city is carried out in a step-by-step manner, which means that the timing of policy impact is not uniform (Hong et al., 2023). Based on this, to effectively identify the impact of digital infrastructure construction on urban carbon reduction, this paper regards the exogenous policy impact of the Broadband China demonstration city as a quasi-natural experiment and constructs the following time-varying DID model: Y i t = β 0 + β 1 P o l i c y i t + λ C o n t r o l i t + τ i + γ t + ε i t (1) where Y _it is the explained variable, indicating the regional carbon emissions and intensity of city i in year t; Policy _it is the core explanatory variable, indicating whether the sample is selected as a Broadband China demonstration city. If the sample is selected, the value of this sample is 1 in this year and subsequent years. Otherwise, the value is 0. Its coefficient β1 represents the impact of digital infrastructure construction on urban carbon reduction, which is the focus of this study. Control _i,t represents a group of control variables, which include per capita GDP, population density, employment structure, level of foreign direct investment, green environment, and level of scientific input by the government. τ _i indicates the city fixed effect. γ _t indicates the year fixed effect. Ɛ _it represents the random perturbation term.

Table 2 shows the results of the baseline regression. Columns (1) and (2) give the estimated results of the effects of digital infrastructure construction on carbon emissions while (3) and (4) give those on carbon intensity. Columns (1, 3) and (2, 4) are respectively without and with the control variables. According to Table 2, when control variables are not considered, only time fixed effects and city fixed effects are controlled, and the Policy coefficient affecting carbon emissions and carbon intensity is significantly negative at the 1% level. When control variables are added, the Policy coefficient remains significantly negative at the 1% level, with coefficient values of −0.029 and −0.057, respectively. This indicates that the construction of digital infrastructure represented by the Broadband China strategy can significantly promote urban carbon reduction. Hong et al. (2023) reached similar conclusions. Among the control variables, The coefficient of x1 is significantly positive, indicating that the higher the level of urban economic development, the greater the urban carbon emissions. This conclusion is similar to the findings of Wang et al. (2024). The coefficient of x2 is significantly negative, indicating that the higher the urban population density, the more concentrated the urban layout and compact development, which reduces the dispersed consumption of energy and resources and reduces urban carbon emissions. The coefficient of x3 is significantly negative, indicating a higher proportion of employment in the service industry, lower dependence on energy, and greater reliance on human capital and information technology, thereby reducing the city’s carbon emissions and intensity. The coefficient of x4 is significantly positive, indicating that the more foreign direct investment a city introduces, the greater its carbon emissions, which to some extent verifies the “pollution shelter hypothesis.” The coefficient of x5 is significantly negative, indicating that the higher the level of urban greening, the more it helps to reduce the carbon content in the air, thereby achieving the effect of emission reduction. The coefficient of x6 is not significant, indicating that the level of urban science expenditure has little impact on urban carbon reduction.

To ensure the accuracy of the results of the DID method and overcome the systematic difference of changing trends between the cities in the experimental and control groups, we used the propensity score matching (PSM) method to test the robustness of the baseline regression. Also, to avoid sample size loss and alleviate selectivity bias as much as possible, we applied adjacent one-to-one matching. In Table 3, Columns (1) and (2) are the regression results of the PSM-DID method for carbon emissions and intensity. The regression results show that the Policy coefficients are significantly negative at the levels of 1%, which indicates that the Broadband China strategy for digital infrastructure construction has reduced carbon emissions and intensity, thereby proving that the benchmark regression results were robust.

In Table 9, Columns (1) is the estimated results of the green technology innovationmechanism. The coefficient of Policy in Column (1) is significantly positive at the level of 1%, indicating that digital infrastructure construction has a significant positive effect on urban green technology innovation, that is, digital infrastructure construction promotes the improvement of urban green technology innovation level. On the one hand, the construction of digital infrastructure means the gathering of high-tech talents such as digital technology, enhancing knowledge spillover, and improving the level of urban green technology innovation. The construction of digital infrastructure can improve urban energy efficiency, optimize resource allocation, and promote the improvement of urban green technology level (Wang et al., 2024). On the other hand, green technology innovation promotes the energy revolution and accelerates the development of clean energy, helping to transform and upgrade the energy consumption structure (Guo et al., 2022), which is conducive to building a clean and low-carbon modern energy system. The improvement of green technology innovation level can promote the green and low-carbon transformation of energy, increase the proportion of clean energy consumption, promote the construction of new energy systems, and continuously reduce energy consumption and urban carbon emissions. Therefore, green technology innovation is an effective means to solve environmental pollution problems (Acemoglu et al., 2012; Guo et al., 2022), and digital infrastructure construction can significantly reduce carbon emissions by improving the level of green technology innovation (Yuan et al., 2022).

In Table 9, Columns (2) is the estimated results of the industrial structure upgrading mechanism. The coefficient of Policy in Column (2) is significantly positive at the level of 10%, indicating that the construction of digital infrastructure has a positive effect on the upgrading of urban industrial structure, that is, the construction of digital infrastructure promotes the improvement of urban industrial structure upgrading. On the one hand, digital infrastructure construction can promote the supply efficiency of enterprise digital elements, improve the integration efficiency of various element resources, and thus give birth to new formats and industries, promoting the optimization of urban industrial structure (Yi et al., 2022). The construction of digital infrastructure empowers urban digital transformation, promotes the integration of digital economy and traditional industries, accelerates urban digital industrialization and industrial digitalization construction, promotes the digital transformation and upgrading of traditional industries, and ultimately achieves the optimization and upgrading of urban industries (Su et al., 2021). On the other hand, upgrading the industrial structure can eliminate outdated equipment, transform the extensive economic development model of cities, and reduce the carbon emissions of high energy consuming and high polluting enterprises (Wu et al., 2021). Upgrading the industrial structure can promote the transformation of urban industries towards high technology and high added value, promote the development of low-carbon and clean industries, and significantly curb carbon emissions (Zhao et al., 2022). Therefore, the construction of digital infrastructure can reduce urban carbon emissions by promoting industrial structure upgrading.

To test for regional heterogeneity in the carbon emission reduction effects of digital infrastructure construction, we divided the 283 sample cities into an eastern regional group but the central and western regions together into another group, then introduced the interaction term of the cities and the Policy variable in each regional group into Equation 1. In Table 10, Columns (1) and (2) are the regression results of the effects of digital infrastructure construction development on carbon emissions and intensity for the eastern region while (3) and (4) are for the central and western ones. The carbon reduction effect of digital infrastructure construction is more evident in the eastern region, but the carbon reduction effect on the central and western regions is only reflected in carbon intensity, and the carbon emissions are not significant. The main reason may be that compared with the central and western regions, the eastern region has a high level of economic development, a good foundation for Internet development, flat terrain and convenient transportation, so the implementation of the Broadband China strategy has promoted the development of urban digital technology, and thus has a stronger effect on urban carbon emission reduction.

We also classified the sample cities according to their population sizes by following the hierarchy criteria of the Outline of the National New Urbanization Plan issued by the State Council. Cities with more than 1 million urban residents are large cities; otherwise, they are small or medium-sized. An interaction term between the Policy variable and a dummy variable for city size was introduced into Equation 1 to test if city-size heterogeneity played a role in the carbon emission reduction effects of digital infrastructure development. In Table 11, Columns (1) and (2) show the regression results for the large cities, whereas (3) and (4) show those for the small and medium-sized ones. The regression results show that the carbon reduction effect of digital infrastructure construction in large cities is significant, while the carbon reduction effect in small and medium-sized cities is not significant. The possible reason is that large cities often have more advantages in Internet development foundation, financial strength and technology level, while the implementation of the Broadband China strategy plays a stronger role in promoting the upgrading of industrial structure and improving the ability of digital technology innovation, and the greater the impact on urban carbon emission reduction. However, the spatial agglomeration of carbon emission entities in small and medium-sized cities, especially enterprise entities, has not yet been fully included in the scope of emission control.

To further analyze the differences in the carbon emission reduction effects of digital infrastructure construction in cities with different levels of economic development, Yao et al. (2017) applied the unchanged control variables and all other model settings by taking 50% as the median per capita GDP. The sample cities were divided into those above and below the median as respectively having high and low levels of economic development. In Table 12, Columns (1) and (2) represent the effects of digital infrastructure construction development in the high-level cities, whereas Columns (3) and (4) represent those in the low-level ones. The former group has seen more significant carbon emission reduction effects due to better digital infrastructure and stronger financial strength. The development of digital infrastructure construction can improve the rationalization and upgrading of urban industrial structure while having more significant inhibitive effects on urban pollution and emission reduction.

Given the close relationship between urban carbon emissions and environmental regulations, and considering the varying intensity of environmental regulations across cities, the impact of digital infrastructure development on urban carbon emissions may exhibit heterogeneity. Accordingly, this study adopts the threshold model proposed by Yuan and Xie (2014) to estimate the annual average intensity of environmental regulations for the period 2006–2023. Based on these annual averages, cities are classified into three groups: strong, moderate, and weak regulation intensity. The results of the group-wise regression analysis, as presented in Table 13, reveal that digital infrastructure construction exerts a more pronounced carbon reduction effect in cities with higher environmental regulation intensity. This can be attributed to two primary mechanisms. First, stronger environmental regulation reflects a greater governmental commitment to addressing carbon emissions and environmental pollution, leading to more direct and stringent regulatory interventions. Second, stringent environmental regulations induce an “innovation compensation” effect, which incentivizes firms to enhance their green technology innovation capabilities, thereby generating a green Porter effect. This effect facilitates the production of cleaner products and contributes to carbon emission reduction. Consequently, the Broadband China pilot policy demonstrates a more significant carbon abatement effect in cities with stronger environmental regulation intensity.