New Infrastructure Construction, Institutional Pressure, and Sustainable Development Performance: Empirical Evidence from Chinese Manufacturing Enterprises

Abstract

The rapid development of new infrastructure has profoundly influenced the development pattern of enterprises. Also, it provides new opportunities for manufacturing enterprises to achieve greater sustainable development performance (SDP). Based on data on manufacturing enterprises from 2012 to 2022, this study combines financial performance and ESG performance to measure the SDP of enterprises and discusses the mechanism through which new infrastructure construction (NIC) affects SDP. The results indicate that NIC significantly promotes the financial performance and ESG performance of enterprises, thereby promoting their SDP. This conclusion remains robust after rigorous testing for endogeneity and other robustness checks. Mechanism analysis reveals that appropriate environmental regulation and market competition can strengthen this positive effect of NIC on the SDP of enterprises, while media attention weakens it. Heterogeneity analysis shows that integration infrastructure in NIC most significantly promotes the SDP of enterprises at the infrastructure level. At the regional level, NIC significantly promotes the SDP of enterprises in eastern and central areas. At the enterprise level, NIC more effectively promotes SDP in state-owned enterprises and in growth periods of enterprises. This study provides a theoretical reference and empirical evidence for enhancing global micro-level sustainable development.

1. Introduction

Enterprises operate within a continuously changing economic and social environment, facing increasing pressures related to survival and development. How to endow enterprises with enduring vitality is a critical question for contemporary entrepreneurs and researchers. In the past, many Chinese manufacturing enterprises primarily prioritized short-term gains, neglecting long-term environmental and social impacts, which posed severe challenges to their sustainable development. On the one hand, cost advantages are unsustainable. Since 2010, China’s manufacturing output has ranked first in the world, yet cost advantages alone cannot ensure sustainable development; breakthroughs in technological and institutional innovation are urgently needed. On the other hand, transformation and upgrading have become imperative under new circumstances. As risks from climate change and environmental degradation escalate, and the digital economy rapidly develops, manufacturing enterprises face the dilemma that traditional production factors can no longer sustain development, creating an urgent need for new drivers to enhance SDP. Therefore, against the historical backdrop of accelerating global technological revolution and intersecting sustainable development strategic goals, how enterprises can adapt and transform their competitive advantages has become a critical challenge for Chinese manufacturing enterprises.

With the latest round of technological revolution and industrial transformation, new infrastructure construction (hereinafter referred to as NIC), represented by 5G communications, big data centers, and artificial intelligence, is catalyzing profound changes in industry and innovation. China has successively introduced NIC construction plans. According to China’s 2024 Communications Industry Statistical Bulletin, China’s digital infrastructure technology is leading the world. As of 2024, the number of 5G base stations in China exceeded 2 million, accounting for over 60% of the global world and covering more than 98% of administrative villages. Furthermore, through the construction of industrial internet, artificial intelligence, and other platforms, a large-scale industrial cluster has been formed. For example, the Xiongan New Area has realized unmanned construction collaborative operations through Beidou + 5G technology, increasing construction efficiency by 40%.

Unlike traditional infrastructure, NIC relies on next-generation information technologies, effectively breaking down information, knowledge, and spatial barriers between production sectors [1]. This transformation may reshape corporate production and green innovation systems [2], thereby impacting SDP. Extensive research has explored the economic effects of NIC. At the enterprise level, studies have analyzed the impact of NIC on export resilience [3], enterprise upgrading [4], total factor productivity [5], and digital transformation [6]. Some studies also examine NIC’s environmental effects from a green transformation perspective. For example, Wang et al. (2024) [7] found that NIC significantly enhanced corporate economic, green innovation, and environmental performance, facilitating green transformation through strengthened internal capabilities, increasing external market attention, and deepened green finance [8]. Gong and Wu (2025) [9] demonstrated that ultra-high-voltage transmission significantly reduces corporate pollution emissions and promotes total factor productivity, thereby promoting green transformation. Gao and Lu (2023) [10] found that digital infrastructure and integration infrastructure can effectively promote both the quantity and quality of corporate green technology innovation, while innovation infrastructure significantly inhibits it. Moreover, industrial agglomeration significantly moderates the relationship between NIC and green technology innovation.

The literature on the antecedents of corporate SDP spans macro, meso, and micro dimensions. At the macro level, factors include climate change, low-carbon policies [11], and green finance policies [12]. At the micro level, factors include corporate environmental ethics [13], green technology innovation [14], and digital transformation [15]. The mechanism through which NIC influences SDP primarily manifests as “stock optimization” and “increment supplementation”. “Stock optimization” refers to NIC acting as a technology spillover, breaking development barriers caused by path dependency on traditional infrastructure, thereby enhancing production efficiency and quality and providing the necessary conditions for promoting enterprise sustainability [16]. “Increment supplementation” refers to NIC’s investment support and innovative development of application scenarios for emerging enterprises. NIC investment effectively absorbs financial resources from the market and converts them into effective supply, laying a solid foundation for enterprise sustainable development. Furthermore, supported by intelligent technologies like AI, big data, internet, and blockchain, NIC can form business incentives and build intelligent industrial ecosystems, thereby promoting SDP [17,18]. Additionally, according to institutional theory, enterprises are embedded within institutional environments, and their innovation behaviors are jointly shaped by contextual factors [19]. Therefore, the relationship between NIC and SDP may be influenced by external institutional pressures; however, few scholars have explored these pressures as potential these pressures as potential contingency factors. To secure essential support and resources from the external institutional environment, enterprises tend to adopt strategies and behaviors that align with the expectations of external stakeholders, thereby complying with institutional demands to achieve organizational legitimacy [20]. In this sense, external institutional pressure serves as a critical external driver that encourages enterprises to channel NIC investments into sustainability-oriented practices.

Existing studies provide a solid foundation but leave several research gaps open for further investigation. Firstly, limited research has directly examined the impact of NIC on the SDP of manufacturing enterprises. With regard to indicator selection, most of the prior literature focuses on specific sub-sectors of NIC (such as digital infrastructure or ultra-high voltage transmission). Studies that do consider NIC holistically often approach it from the perspective of investment stock, overlooking its broader role as a techno-economic enabler. Furthermore, the concept of environmental, social, and governance (ESG) criteria is inherently aligned with sustainable development at the macro leve and has become a key metric for assessing corporate sustainability. Nevertheless, current evaluations corporate SDP predominantly emphasize economic, social, and environmental outcomes individually, often neglecting the central role of ESG in defining sustainable performance. Comprehensive assessments integrating both financial performance and ESG metrics remain underdeveloped. While extensive research has focused on the economic implications of NIC, the sustainability value it generates has received relatively little attention. Given the rapid advancement of digital technologies and increasing corporate emphasis on sustainability [21], there is a pressing need to thoroughly examine the relationship, underlying mechanisms, and heterogeneous effects between NIC and SDP.

This study measures enterprise-level SDP through a composite lens combining financial performance and ESG performance, constructs a comprehensive evaluation index for NIC, and systematically investigates the impact and mechanisms through which NIC influences the SDP of manufacturing enterprises. Relative to the existing literature, this study makes three marginal contributions: (1) By focusing on manufacturing enterprise sustainability—rather than solely economic outcomes—it integrates financial and ESG performance indicators, thereby expanding the evaluative dimensions of SDP and enriching scholarly understanding of the interplay between economic performance and ESG outcomes. (2) Grounded in institutional theory, this study incorporates institutional pressures into the analytical framework, providing both theoretical justification and empirical validation of the moderating roles played by environmental regulation, market competition, and media scrutiny in the NIC and SDP relationship. This extends the applicability of institutional theory to the context of new infrastructure development. (3) This study explores the heterogeneous effects of NIC on SDP across enterprise, industry, and regional dimensions, contributing nuanced insights and advancing knowledge in this domain.

2. Theoretical Analysis and Research Hypotheses

2.1. The Direct Impact of NIC on SDP

This study posits that NIC contributes to enhancing the SDP of manufacturing enterprises. In pursuing sustainability, enterprises not only need to focus on economic benefits but also balance environmental and social considerations to achieve harmonious coexistence with their surrounding environments. Building on Hasan et al. (2023) [22], this study conceptualizes SDP as comprising two core dimensions: financial performance and ESG performance.

Regarding to financial performance, NIC facilitates the reconfiguration of production factors and enhances capacity utilization, thereby improving financial outcomes. Through the advancement and integration of next-generation information technologies, NIC permeates all stages of the manufacturing industry chain [23]. Firstly, NIC enables more precise matching between upstream and downstream enterprises via supply chain digitalization, reducing excess capacity and improving capacity and optimizing resource allocation. By advancing reforms in digital technologies such as 5G, the Internet of Things (IoT) and big data analytics, NIC expands technological frontiers in production and fosters the emergence of new growth drivers [24]. Secondly, NIC drives intelligent transformation across the entire manufacturing process. The large-scale deployment of NIC spurs changes throughout traditional business models, accelerating the upgrading of conventional production methods and extending value-added segments along the value chain [25]. Thirdly, NIC improves operational management efficiency. For instance, big data analytics enable enterprises to forecast market demand more accurately, optimize inventory levels, and reduce overproduction. IoT systems support real-time monitoring of supply chains, enhancing the self-organization and precision of production processes, thus contributing to improved financial performance [26].

Regarding to ESG performance, firstly, NIC plays a significant enabling role. First, it enhances transparency and traceability in resource utilization and pollutant emissions [27]. Digital technologies such as IoT and cloud computing allow for dynamic tracking and real-time analysis of material flows and industrial waste. By precisely monitoring energy and resource consumption and optimizing end-to-end processes, NIC helps minimize unnecessary waste and environmental pollution [28]. Second, NIC supports intelligent energy allocation, improves production efficiency, and enhances resource integration, thereby mitigating the environmental impact of manufacturing activities and strengthening environmental governance capabilities [29]. Third, NIC encourages greater corporate engagement in environmental, social, and governance (ESG) initiatives—an investment approach that seeks to generate not only financial returns but also positive social and environmental impacts. Enterprises can leverage cutting-edge technologies such as artificial intelligence (AI), blockchain, and virtual reality (VR) to strengthen ESG monitoring mechanisms, enhance learning and adaptation, and promote robust ESG disclosure systems, thereby advancing overall SDP [30]. Based on the above theoretical analysis, this study proposes the following hypotheses:

Hypothesis 1.

NIC positively enhances the SDP of manufacturing enterprises.

Hypothesis 1a.

NIC positively enhances the financial performance of manufacturing enterprises.

Hypothesis 1b.

NIC positively enhances the ESG performance of manufacturing enterprises.

2.2. The Moderating Effect of Institutional Pressure

SDP is influenced by the institutional pressures that enterprises face, which are primarily categorized into regulative, normative, and mimetic pressures. By conforming to these institutional pressures, enterprises acquire organizational legitimacy, an essential external resource that facilitates social recognition and access to support, thereby significantly enhancing their SDP [31].

Regulative institutional pressure refers to the coercive influence exerted by governmental and regulatory authorities through laws, regulations, and policy instruments to monitor and shape organizational behavior. In the context of enterprise SDP, this pressure manifests in the form of environmental regulation, including legislative mandates, incentive-punishment mechanisms, and pollution control targets. Therefore, this study employs the intensity of environmental regulation as a proxy for regulative institutional pressure.

Although stringent environmental regulations may increase compliance costs and divert financial resources from core operations, strategic integration of policy tools, such as carbon emission trading systems and government subsidies, can mitigate these burdens. Such mechanisms encourage investment in advanced technologies with higher resource efficiency and improved environmental performance, thereby enhancing enterprise SDP [32]. Environmental regulation imposes strict constraints on corporate production and operational activities. To enhance regulatory legitimacy, minimize political risks, and secure greater access to resources and long-term viability, enterprises are incentivized to channel digital resources into sustainability initiatives, actively aligning their practices with regulatory expectations [33]. Under the guidance of environmental regulation, enterprises are more likely to leverage the technological advantages of new infrastructure construction (NIC) to develop and deploy green technologies, promote industrial restructuring, and achieve deep integration between NIC and green innovation [34]. Furthermore, driven by profit maximization objectives and the pursuit of future market opportunities, environmental regulation compels enterprises to strategically deploy NIC to reduce pollutant emissions at the source, lower penalties for non-compliance, and gain competitive advantage in increasingly regulated markets [35]. Based on this analysis, the following hypothesis is proposed:

Hypothesis 2.

Environmental regulation positively moderates the relationship between NIC and the SDP of manufacturing enterprises.

Mimetic pressure arises from inter-enterprise competition, reflecting the tendency of organizations to imitate the behaviors of industry peers to maintain perceived legitimacy and competitiveness. Consequently, when evaluating NIC investments, enterprises incorporate considerations of market competition into their strategic decision making, which in turn influences the impact of NIC on SDP. Thus, this study uses market competition as an indicator of mimetic institutional pressure.

Market competition motivates enterprises to pursue differentiated competitive advantages. As ecological transformation accelerates and customer preferences shift toward sustainable products, both cooperative and competitive dynamics within the manufacturing sector intensify [36]. In such an environment, while enterprises share common goals of ecological modernization and stakeholder engagement, rivalry among them grows stronger. Under high competitive pressure, enterprises are more inclined to engage in green innovation races to achieve sustainability leadership [37]. The advanced technologies and digital infrastructure enabled by NIC provide critical enablers for implementing sustainability-oriented strategies. To differentiate themselves in the marketplace, enterprises are likely to intensify their utilization of NIC capabilities and increase, investment in green and sustainable applications of new infrastructure [38]. For instance, leveraging the supported by industrial internet, enterprises can achieve intelligent management of production processes, reducing energy consumption and emissions, thereby improving product-level green competitiveness [39]. Moreover, heightened market competition increases enterprises’ willingness to adopt digital and intelligent technologies. Within the NIC ecosystem, new standards and competitive rules emerge, and intensified rivalry amplifies environmental uncertainty. To reduce information asymmetry, secure first-mover advantages in sustainability practices, and strengthen green competitiveness, enterprises are more inclined to harness NIC to advance their sustainability agendas [40]. Based on this theoretical reasoning, the following hypothesis is proposed:

Hypothesis 3.

Market competition positively moderates the relationship between NIC and the SDP of manufacturing enterprises.

Normative pressure constitutes an informal component of the institutional environment, primarily encompassing socially embedded values and norms that shape enterprises’ behavior through moral and ethical expectations [41]. In this study, media attention is employed as a proxy for normative institutional pressure. Prior research indicates that media attention exerts dual effects, potentially generating two distinct types of influence: on one hand, it serves as an effective channel for enterprises to understand stakeholder expectations and sustainability-related concepts, thereby enhancing corporate focus on sustainable development through governance mechanisms; on the other hand, excessive media scrutiny may constrain organizational autonomy, amplify perceived risks of sustainability initiatives, and ultimately undermine support for green technology innovation [42]. Therefore, the moderating role of media attention between NIC and SDP warrants further empirical investigation.

This study posits that media attention can impair the image and managerial reputation of enterprises, thereby weakening their motivation to pursue sustainability practices through NIC. Grounded in the market pressure hypothesis, although media coverage can enhance enterprises’ accountability, it simultaneously generates substantial external pressure that may induce managerial short-termism, leading firms to forgo high-risk but high-potential-return investments in innovation. This effect is particularly pronounced for heavily polluting industries, where media coverage tends to be negative and may result in reputational damage, investor skepticism, and constraints on long-term development. Under intense media scrutiny, enterprises face mounting pressure to demonstrate progress in digital transformation and green innovation. When actual innovation outcomes fall short of expectations or short-term financial performance declines, enterprises may encounter amplified criticism and public distrust, and these effects that are often magnified through rapid information dissemination, potentially triggering sharp declines in stock prices [43]. Such pressures may lead to overly conservative strategic decisions, reluctance to adopt emerging technologies, and, in extreme cases, “greenwashing” behaviors, including the fabrication of environmental achievements or the superficial embellishment of corporate image [44].

Moreover, frequent media engagement may distract management and disrupt normal operational and innovation processes. Enterprises are often required to dedicate considerable time and resources to respond to media inquiries, prepare press releases, conduct interviews, and manage public relations campaigns related to corporate social responsibility [45], diverting attention and capital away from core R&D and operational activities. Additionally, media reporting may inadvertently disclose sensitive information about ongoing innovation projects, exposing firms to competitive risks and impeding the successful implementation of NIC-driven initiatives [46]. Based on this theoretical analysis, the following hypothesis is proposed:

Hypothesis 4.

Media attention negatively moderates the relationship between NIC and the SDP of manufacturing enterprises.

3. Research Design

3.1. Data Source and Processing

Given that China formally introduced proposed the “Five-in-One” overall strategic layout in 2012, integrating ecological civilization construction into the framework of socialism with Chinese characteristics, the temporal scope of this study commences in 2012. With respect to enterprises, manufacturing enterprises continue to exhibit characteristics of high input, high consumption, and high pollution. In particular, heavily polluting manufacturing enterprises represent a primary focus for environmental regulation and policy intervention. Moreover, traditional heavily polluting industries are increasingly reliant on advanced digital technologies such as artificial intelligence (AI) and the industrial internet to advance green transformation and intelligent upgrading. This creates significant potential for deep integration between NIC and these sectors. Therefore, this study focuses on heavily polluting manufacturing industries as the primary research population. Furthermore, drawing upon the List of Industries Subject to Environmental Verification for Listed Companies (2008), the Guidelines for Industry Classification of Listed Companies (2012), and building on the classification approach employed by Xi and Zhao (2022) [47], this study identifies the following manufacturing sectors as heavily polluting enterprises (see

Table 1. Heavy pollution manufacturing industry segments and its industry code.

Heavily Polluting Manufacturing Sub-Sector	Two-Digit Industry Code(s)	Heavily Polluting Manufacturing Sub-Sector	Two-Digit Industry Code(s)
Textiles/Fur/Leather	C17 Textiles	Petrochemicals/ Plastics	C25 Petroleum Processing, Coking & Nuclear Fuel Processing
	C18 Apparel& Accessories		C26 Chemical Raw Materials & Products
	C19 Leather, Fur, Feather & Related Products, Footwear		C29 Rubber & Plastics
	C28 Chemical Fibers	pharmaceutical manufacturing	C27 Pharmaceuticals
Metals/Non-Metals	C30 Non-metallic Mineral Products	Automotive Manufacturing	C36 Automotive
	C31 Smelting & Pressing of Ferrous Metals
	C32 Smelting & Pressing of Non-ferrous Metals	Light Industry	C15 Beverages
	C33 Metal Products		C22 Paper & Paper Products

).

To maintain the integrity and consistency of the data, missing data (such as the capacity of the bureau switch, the number of technology business incubators) were supplemented using interpolation to complete the data, The relevant variables passed the sensitivity analysis. The data come from the National Bureau of Statistics, the China Statistical Yearbook, China Science and Technology Statistical Yearbook, the China High-Tech Industry Statistical Yearbook, the China Electronic Information Industry Statistical Yearbook, the China Torch Statistical Yearbook and other yearbooks. Moreover, in order to avoid the interference of outliers and too many missing values, removing the Tibetan provinces that are special in politics and economy, and examine the NIC level of 30 provinces in mainland China (excluding Tibet).

The enterprise-level data are sourced from A-share listed manufacturing enterprises on the Shanghai and Shenzhen Stock Exchanges. To ensure data reliability and completeness, the sample was constructed through a multi-step screening process: first, 5364 listed companies were initially identified based on the manufacturing sector classification in the Wind database; second, 883 enterprises belonging to heavily polluting manufacturing industries were selected according to established industry classifications; third, ST and PT-designated enterprises, indicating special treatment for financial distress, as well as companies without publicly available ESG reports during the period 2012–2022, were excluded. The final sample consists of a panel dataset with 1881 observations from 171 listed heavily polluting manufacturing enterprises over the 11-year period from 2012 to 2022. Missing values in the dataset were addressed using regression-based imputation techniques. Financial performance indicators and control variables were obtained from the CSMAR database. ESG performance scores were derived from the Hua Zheng ESG ratings provided by the Wind database. Data on environmental regulation, market competition, and media attention were extracted manually through textual analysis of corporate annual reports and provincial government work reports, ensuring alignment with institutional context and policy dynamics.

3.2. Description of Variables

3.2.1. Explained Variable

Building on established scholarly work, this study operationalizes the SDP of manufacturing enterprises as a dual-dimensional framework encompassing financial performance and ESG performance. Financial performance is measured by Return on Assets (ROA), a widely recognized indicator of profitability that reflects a enterprise’ efficiency in generating earnings from its asset base [48]. ESG performance is captured using enterprise-level ESG scores derived from Wind ESG ratings, which developed by Shanghai Huazheng Index Information Service Company (Shanghai, China). Consistent with prior studies [49], the Huazheng ESG rating system comprises nine tiers, where higher scores indicate stronger performance across environmental stewardship, social responsibility, and corporate governance dimensions. The standardization procedure applied to these variables is formally specified in Equations (1) and (2).

Firstly, financial performance and ESG performance are standardized to 0–1 range using:

y* = (y − miny)/(maxy − miny)

(1)

where miny and maxy denote the minimum and maximum values of the original data of financial performance and ESG performance data, respectively, and y* represents the corresponding normalized value.

Secondly, drawing on organizational ambidexterity theory, which emphasizes the strategic balance and simultaneous pursuit of financial performance and environmental/social responsibility under conditions of resource constraints. Thus, the following formula is used to calculate SDP.

$$Sdp=\left[\left(1-\left|Fin-Esg\right|\right)\times \sqrt{Fin\times Esg}\right]/1$$

(2)

where Fin, Esg, Sdp indicate the financial performance, ESG performance and the sustainable development performance after normalized, respectively.

3.2.2. Explanatory Variable

This study conceptualizes NIC as a techno-functional-service complex that constitutes the foundational spaces for economic and social activities, extending beyond mere technical facilities to encompass integrated techno-economic systems. Drawing on the classification framework of NIC established by China’s National Development and Reform Commission, and guided by the principles of scientific rigor, representativeness, and data availability, an evaluation system was developed to comprehensively capture the core domains of NIC within the tripartite structure of information infrastructure, integration infrastructure, and innovation infrastructure. Indicators with significant data limitations—such as those related to the industrial internet and new energy vehicle (NEV) charging piles—were excluded to ensure measurement reliability. The entropy method was employed to determine indicator weights, resulting in a 27-indicator composite system for assessing the development level of NIC in China [50], as presented in

Table 2. New infrastructure development evaluation index system.

System Level	Criterion Level	Primary Indicator Level	Secondary Index Layer	Weight
The development level of NIC	Information infrastructure	Technology utilization	X1: Mobile subscription (set/hundred person)	0.0086
			X2: Length of cable line (km/102 square kilometers)	0.0163
			X3: Local switch capacity (million gates)	0.0352
			X4: Internet access port (104 piece)	0.0203
			X5: Number of internet broadband access users (million households)	0.0220
			X6: Number of domain names (piece/104 person)	0.0553
		Economic benefit	X7: Total volume of post and telecommunications business (billion yuan)	0.0429
			X8: Information technology service income (million yuan)	0.0681
			X9: Software product revenue (million yuan)	0.0616
	Integration infrastructure	Integration format	X10: Number of public buses and trams in operation (vehicle)	0.0214
			X11: Number of digital TV users (thousand households)	0.0177
			X12: Number of automatic weather station sites (piece)	0.0160
			X13: The proportion of green energy power generation (%)	0.0239
			X14: Express quantity (104 piece)	0.0788
		Information degree	X15: The number of computers per 100 people (set)	0.0136
			X16: Number of websites per 100 enterprises (piece)	0.0036
			X17: Software business income (million yuan)	0.0654
			X18: E-commerce sales (billion yuan)	0.0488
	Innovation infrastructure	R&D strength	X19: Proportion of R&D personnel (%)	0.0346
			X20: Intensity of R&D expenditure input (%)	0.0210
			X21: Number of R&D institutions in high-tech industry (item)	0.0786
			X22: Funding for the development of new products (million yuan)	0.0497
			X23: Patent application number (Piece)	0.0548
		Science and education service	X24: The number of visitors to the science and technology museum that year (million person)	0.0230
			X25: Number of valid domestic patents (item)	0.0529
			X26: Number of R&D projects in universities (item)	0.0204
			X27: Number of technology business incubators (piece)	0.0456

.

Information Infrastructure. This domain encompasses network infrastructure, computing power infrastructure, and new technology infrastructure. Rooted in next-generation information technologies, it supports widespread technological application and delivers economic benefits across diverse sectors of daily life and industrial production.

Integration Infrastructure. This refers to infrastructure forms emerging from the deep integration of information technology into traditional physical infrastructure, enabling intelligent transformation. From a techno-economic paradigm perspective, it reflects the convergence of infrastructure systems and the advancement of digitalization across industrial sectors.

Innovation Infrastructure. Originating from scientific research, technological development, and product R&D activities, this domain focuses on major science and technology infrastructures, educational facilities, and public research platforms, characterized by their public welfare orientation. Based on its conceptual foundation and structural classification, two sub-dimensions are identified: R&D intensity and education and science services.

3.2.3. Moderating Variable

Environmental regulation (Ere). Given the diversity of China’s environmental regulatory instruments and the programmatic function of government work reports as official summaries of policy priorities and administrative decisions, this study employs the frequency (or proportion) of environment-related terms in the Reports on the Work of the Government—sourced from the China Government Work Report Database (CGWRD)—as a proxy for governmental environmental regulation stringency. This indicator is presented in

Table 3. Terms related to environmental regulation.

Core Word

Segmentation Dictionary

environmental regulation

Environmental protection, environmental protection, haze, haze control, pollution, energy consumption, pollution control, emission reduction, sewage discharge, ecological environment, ecological protection, ecological destruction, water ecology, low carbon, sulfur dioxide, carbon dioxide, PM10, PM2.5, chemical oxygen demand, COD, scattered pollution, emissions, air, water environment, water quality, blue water, black odor, sewage, waste gas, waste residue, environmental violations, environmental crimes, environmental cases, environmental penalties, environmental governance, environmental quality, blue sky, coal burn, green, dust, and exhaust gas.

. Following the methodology of Chen and Chen (2018) [51], Python 3.8 was used to calculate the ratio of the number of words associated with environmental protection words in each regional report to the total words count, thereby quantifying the intensity of environmental regulation.

Market competition (Cre). The Herfindahl–Hirschman Index (HHI) combines the advantages of both absolute and relative measures of market concentration and is less sensitive to variations in enterprise size distribution and the number of enterprises in the market, making it a robust indicator of market competitiveness. Accordingly, this study calculates the uses HHI based on enterprise asset size to measure of the degree of market competition within industries. A higher HHI value indicates lower market competition (i.e., greater market concentration), while a lower value reflects more intense competition [52].

Media attention (Med). Drawing on established research, media attention is measured by the volume of media coverage related to individual enterprises. Data were extracted from the China Core Newspapers Database (CCND) through a systematic process involving keyword searches and manual screening to ensure accuracy and relevance. The number of news articles reporting on each enterprise during the study period is used as the operational measure of media attention [53].

3.2.4. Control Variable

Given the complexity of factors influencing SDP, and to enhance measurement accuracy and minimize model specification errors, this study incorporates control variables that capture key financial characteristics and organizational governance attributes, drawing on the established literature [54]. These control variables include enterprise scale (Siz), measured as the lnatural logarithm of year-end total assets, reflecting the scale of operations; financial leverage (Dfl), defined as the ratio of the percentage change in earnings per common share to the percentage change in earnings before taxes, indicating financial risk exposure; capital expenditure (Cae), calculated as the ratio of cash outflows for the acquisition of fixed assets, intangible assets, and other long-term assets to total, representing investment intensity; Tobin’s Q (TbQ), computed as the ratio of the enterprise’s market value to its total assets, serving as a proxy for growth opportunities and market valuation; board size (Boa), measured by the total number of board members, reflecting governance structure complexity; proportion of independent directors (Ind), expressed as the ratio of the number of independent directors to the total number of board members, indicating board independence; executives duality (Dua), captured using a dummy variables: assigned a value of 1 if the positions of chairman and general manager are held by the same individual, and 0 otherwise, reflecting potential concentration of decision-making power; institutional ownership (Ins), measured as the proportion of total shares held by institutional investors, representing external monitoring and ownership stability. Definition of variables and data sources of this study are shown in

Table 4. Definition of variables and data sources.

Variable Type	Variable Name	Variable Measurement	Data Sources
Explained variable	Sustainable development performance (Sdp)	Calculated by ROA and ESG	Manual measurement
	Financial performance (Fin)	Net profit/average balance of total assets (ROA)	CSMAR
	ESG performance (Esg)	Enterprise ESG rating	Huazheng ESG
Explanatory variable	New infrastructure construction (Nic)	Composite measurement index	Manual measurement
Moderating variable	Environmental regulation (Ere)	Proportion of environmental protection words in government work report	CGWRD
	Market competition (Cre)	Herfindahl Index	CSMAR
	Media attention (Med)	Number of media reports	CCND
Control variable	Enterprise scale (Siz)	Natural logarithm of total assets	CSMAR
	Financial leverage (Dfl)	The ratio of the change rate of common stock earnings to the change rate of earnings before interest and tax
	Capital expenditure (Cae)	The ratio of cash paid for fixed assets, intangible assets, and other long-term assets to total assets
	Tobin’s Q (TbQ)	Enterprise market value divided by total assets
	Board size (Boa)	Number of board of directors
	Proportion of independent directors (Ind)	Independent directors divided by number of directors
	Executives duality (Dua)	Whether the chairman and general manager are the same person
	institutional ownership (Ins)	Shares held by institutional investors in a company as a proportion of total equity

.

3.3. The Baseline Model

This study links provincial-level data on NIC with firm-level microdata on SDP to examine the impact of NIC on enterprises’ sustainability outcomes. To ensure appropriate model selection, a series of diagnostic tests including F-tests, Lagrange Multiplier (LM) tests, Breusch–Pagan (B–P) tests, and Hausman specification tests were conducted, supporting the adoption of a fixed-effects model for regression analysis. To account for unobservable macro-level influences such as regional policy shifts and fluctuations in market demand, the model incorporates interaction terms between province-industry and time fixed effects. Estimations employ province-clustered robust standard errors to mitigate potential omitted variable bias and enhance the reliability of coefficient estimates. The econometric models are specified as follows:

$$\left\{\begin{array}{c}{Y}_{icjt}=a_0+a_1{Nic}_{ct}+a_2\overline{C}+{\mu }_{cj}+{\sigma }_t+{\epsilon }_{icjt}\\ {Sdp}_{icjt}=b_0+b_1{Nic}_{ct}+b_2{M}_{icjt}+b_k\overline{C}+{\mu }_{ct}+{\sigma }_t+{\epsilon }_{icjt}\\ {Sdp}_{icjt}=c_0+c_1{Nin}_{ct}+c_2{M}_{icjt}+c_3{Nic}_{ct}\times M_{icjt}+b_k\overline{C}+{\mu }_{ct}+{\sigma }_t+{\epsilon }_{icjt}\end{array}\right.$$

(3)

where the subscripts i, c, j, and t denote enterprise, province, industry and year, respectively. Y_icjt represents the dependent variable and encompasses; financial performance (Fin), ESG performance (Esg), and SDP (Sdp_icjt). Nic_ct represents the level of new infrastructure construction. The moderating variable M_icjt captures institutional pressure, including environmental regulation (Ere), market competition (Cre), and media attention (Med). $\overline{C}$ represents a series of control variables. ${\mu }_{cj},{\sigma }_t$ represent province-industry fixed effects and year fixed effects, respectively. ${\epsilon }_{icjt}$ is the random error term.

4. Results and Discussion

4.1. Baseline Results

Table 5. Benchmark regression results.

Variables	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg
Nic	0.0308 *** (7.13)	0.0109 *** (4.71)	0.2854 *** (8.19)	0.0252 *** (6.16)	0.0078 *** (3.75)	0.2367 *** (7.12)
_Cons	0.4377 *** (53.89)	0.0343 *** (7.89)	3.5954 *** (54.78)	−0.2983 *** (−3.17)	−0.2635 *** (−6.17)	−3.1383 *** (−4.18)
Controls	No	No	No	Yes	Yes	Yes
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.3122	0.2875	0.3252	0.3596	0.3945	0.3897

Note: *** represent significant at 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

presents the benchmark regression results. To account for unobserved industry-specific characteristics, macroeconomic fluctuations, and other time-varying confounding factors, both year fixed effects and province-industry interaction fixed effects are included in the model. Additionally, robust standard errors are clustered at the provincial level to address potential heteroskedasticity and within-province correlation. Columns (1)–(3) report the regression results without control variables. without any control variables. The coefficients on Nic are 0.0308, 0.0109 and 0.2854, respectively, all statistically significant at the 1% level, indicating a strong positive association between NIC and SDP. In Columns (4)–(6), after incorporating firm-level control variables, the NIC coefficients remain positive and significant at the 1% level, demonstrating the robustness of the baseline findings. These results consistently show that NIC has a statistically and economically meaningful impact on financial performance and ESG performance, and exerts a significantly positive effect on SDP.

4.2. Robustness Test

4.2.1. Instrumental Variable

This study investigates the impact of NIC on the SDP of manufacturing enterprises at the micro level. Logically, there is no obvious reverse causal problem, as the SDP of individual enterprises will not affect the overall NIC in the province level. However, the empirical research in this study still faces some endogenous challenges: the SDP of enterprises and the development of NIC may be affected by unobservable factors at the provincial level at the same time, so there is a problem of missing variables. In order to solve the above endogenous problems, firstly, this study refers to Huang et al. (2019) [55], and uses the total amount of post and telecommunications business in each province in 1984 as a tool variable for NIC. From the perspective of correlation, postal business can reflect the local logistics facilities construction level, related industry labor and postal income and other factors, representing the predecessor of new infrastructure. From the perspective of exogeneity, the instrumental variable data selected in this study is 1984, which is far from the sample data in this study. At the same time, the postal business has no significant impact on the SDP of enterprises, which meets the exogeneity conditions. Secondly, the “smart city” pilot policy was chosen as an instrumental variable. Therefore, this study draws on the method of Song et al. (2021) [56], uses the “smart city” policy pilot cities as instrumental variables, and uses the IV-2SLS method for empirical analysis.

Table 6. Instrumental variable method.

Variables	Instrumental Variable 1				Instrumental Variable 2
	(1) First Nic	(2) Second Sdp	(3) Second Fin	(4) Second Esg	(5) First Nic	(6) Second Sdp	(7) Second Fin	(8) Second Esg
Nic_IV1	1.6028 *** (10.38)
Nic_IV2					0.5081 *** (6.08)
Nic		0.0321 *** (3.40)	0.0133 *** (4.37)	0.2854 *** (3.89)		0.0849 *** (3.08)	0.0211 *** (3.05)	0.8581 *** (3.98)
_Cons	−2.4643 *** (−6.26)	−0.2093 ** (−2.40)	−0.2559 *** (−6.38)	−2.3394 *** (−3.25)	−0.5655 *** (−3.52)	−0.1568 * (−1.75)	−0.2876 ** (−2.49)	−1.7664 ** (−2.21)
Controls	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2		0.3587	0.3923	0.3890		0.2938	0.3944	0.2799
F		21.20	77.54	215.55		20.09	65.56	76.19
K-P LM statistic		69.112 ***	69.112 ***	69.112 ***		29.254 ***	29.254 ***	29.254 ***
C-D Wald F statistic		562.786 [16.38]	562.786 [16.38]	562.786 [16.38]		74.192 [16.38]	74.192 [16.38]	74.192 [16.38]
Hansen J statistic P		0.2813	0.2813	0.2813		0.2813	0.2813	0.2813

Note: *, **, *** represent significant at 10%, 5%, and 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

shows the results.

The first-stage regression results show that both instrumental variables are strongly correlated with NIC, with statistically significant coefficients. The Kleibergen–Paap LM statistic rejects the null hypothesis of underidentification at the 1% level, confirming that the instruments are identified. The Cragg–Donald Wald F statistic exceeds conventional critical values and is significant at the 1% level, rejecting the presence of weak instruments. Additionally, the overidentification test yields a Hansen J-statistic with a p-value greater than 0.1, indicating that the instruments are jointly valid and satisfy the exclusion restriction. Estimates from both IV models consistently show a significantly positive effect of NIC on enterprise SDP, aligning closely with the baseline regression results. This confirms the robustness of the findings after addressing potential endogeneity.

4.2.2. Replace the Regression Model and Explanatory Variables

To ensure the robustness of the findings, this study conducts additional regression analyses using alternative model specifications and explanatory variables. Firstly, the regression model is replaced with quantile regression, which does not require strict assumptions about the distribution of error terms and is less sensitive to outliers, thereby offering greater estimation robustness To further improve the reliability of coefficient estimates, the Bootstrap method is employed, which approximates the sampling distribution through repeated resampling from the original dataset. In this analysis, a quantile of 0.5 (median) is estimated based on 1000 bootstrap replications. As shown in

Table 7. Replace the regression model and explanatory variables.

Variables	Bootstrap Quantile Regression			Replace Explanatory Variables
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg
Nic	0.0081 *** (4.72)	0.0018 * (1.82)	0.1246 *** (6.52)	0.0747 *** (6.14)	0.0233 *** (3.81)	0.7051 *** (7.25)
_Cons	0.1731 ** (2.23)	−0.1985 *** (−5.56)	−2.1959 *** (−3.20)	−0.2984 *** (−3.17)	−0.2635 *** (−6.14)	−3.1381 *** (−4.17)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.3289	0.3616	0.3681	0.3594	0.3945	0.3898

Note: *, **, *** represent significant at 10%, 5%, and 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

, Columns (1)–(3) show the regression coefficient of Nic on Sdp of manufacturing enterprises is 0.0081, which is significant at 1% confidence level. The regression coefficient of financial performance (Fin) is 0.0018, which is significant at 10% confidence level. The regression coefficient of ESG performance (Esg) is 0.0018, which is significant at 10% confidence level. Therefore, the conclusion that NIC has a significant positive impact on the SDP of manufacturing enterprises is robust. Secondly, change the explanatory variables. In this study, the mean values of information infrastructure, integration infrastructure and innovation infrastructure measured above are substituted into the regression equation as the substitution variables of NIC. Column (4)–Column (6) show that the coefficient of Nic is positive and passes the significance test at the 10% level. This result supports the positive impact of the NIC on the SDP of manufacturing enterprises.

4.2.3. Other Robustness Tests

Considering the outbreak of the novel coronavirus pandemic in late 2019 and its continuation into 2020, which significantly disrupted the manufacturing sector and subsequently affected firms at the micro level, potentially distorting sustainable development trajectories, this study conducts robustness checks to ensure the validity of its empirical findings. First, to mitigate the potential confounding effect of the pandemic, all enterprise observations from 2020 are excluded from the sample. As shown in

Table 8. Other robustness tests.

Variables	Excluding 2020 Samples			Explanatory Variable Lags One Period
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg
Nic	0.0232 *** (5.28)	0.0073 *** (3.25)	0.2092 *** (5.77)	0.0261 *** (5.41)	0.0086 *** (3.74)	0.2716 *** (7.09)
_Cons	−0.2837 *** (−2.87)	−0.2685 *** (−6.04)	−2.9563 *** (−3.85)	−0.3184 *** (−3.33)	−0.2724 *** (−6.05)	−3.2097 *** (−4.10)
Controls	Yes	Yes	Yes	Yes	Yes	Yes
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.3596	0.3906	0.3894	0.3725	0.4034	0.3974

Note: *** represent significant at 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

, Columns (1)–(3) present the regression results after removing the 2020 data. The coefficients on NIC remain significantly positive at the 1% significance level for financial performance (Fin), ESG performance (Esg), and overall SDP, indicating that the core relationship persists even without pandemic-affected periods. Second, to address potential endogeneity arising from reverse causality or simultaneity, a one-period lagged specification of NIC is employed. The corresponding regression results are reported in Columns (4)–(6). At the 1% significance level, lagged NIC exhibits a significantly positive effect on SDP, financial performance, and ESG performance, confirming the directional stability of the estimated impact. In summary, under both alternative sample specifications, excluding 2020 observations and using lagged explanatory variables. The finding that NIC positively enhances the sustainable development performance of manufacturing enterprises remains statistically robust, reinforcing the reliability and consistency of this study’s conclusions.

4.3. Mechanism Test

This study uses Stata17.0 to examine the moderating effect of institutional pressures, including environmental regulation, market competition and media attention. On the relationship between NIC on the SDP of manufacturing enterprises. To mitigate potential multicollinearity arising from interaction terms, all continuous variables are mean centered prior to model estimation. The results of the moderation analysis are presented in

Table 9. The moderating effect of institutional pressure.

Varibles	Sdp
	(1)	(2)	(3)	(4)
Nic	0.0823 *** (3.33)	0.0312 *** (7.17)	0.0258 *** (6.46)	0.0264 *** (6.27)
Ere		0.0352 *** (3.30)
Nic × Ere		0.0227 *** (3.03)
Cre			0.0679 * (2.14)
Nic × Cre			0.0175 *** (3.54)
Med				0.0056 (1.51)
Nic × Med				−0.0032 *** (−3.00)
_Cons	−0.2983 *** (−3.17)	−0.3294 *** (−3.48)	−0.2129 ** (−2.35)	−0.2624 *** (−2.64)
Controls	Yes	Yes	Yes	Yes
Year Fe	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes
R2	0.4380	0.3635	0.3490	0.3623

Note: *, **, *** represent significant at 10%, 5%, and 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

.

4.3.1. The Moderating Effect of Environmental Regulation

As shown in Columns (1) and column (2) of Table 9, the interaction term between new infrastructure construction and environmental regulation (Nic × Ere) has a positive and statistically significant effect on SDP ($\mathsf{\lambda }=0$.0227, p < 0.01 ). This indicates that stricter environmental regulations strengthen the positive impact of NIC on SDP. A plausible explanation is that the green orientation of environmental regulation aligns with the technological characteristics of new infrastructure, guiding enterprises toward adopting energy-saving and emission-reduction technologies embedded in NIC, thereby enhancing their SDP.

4.3.2. The Moderating Effect of Market Competition

According to the results in Columns (1) and (4) of Table 9, the coefficient of the interaction term between new infrastructure and market competition (Nic × Cre) is positive and significant ($\mathsf{\lambda }=0$.0175, p < 0.01), indicating that market competition significantly strengthens the relationship between NIC and SDP. This suggests that intensified market competition incentivizes firms to adopt sustainable practices to attract investors and consumers. As a result, enterprises are more likely to increase investment in and utilization of new infrastructure projects, amplifying its effectiveness in promoting SDP.

4.3.3. The Moderating Effect of Media Attention

The results in Columns (1) and (5) of Table 9 show that the interaction term between new infrastructure construction and media attention (Nic × Med) has a significant negative effect on SDP ($\mathsf{\lambda }=-$0.0032, p < 0.01). This implies that higher levels of media attention weaken the positive impact of NIC on SDP, revealing a negative moderating role. One possible reason is that heightened media scrutiny imposes additional costs on enterprises, such as crisis management and mandatory environmental upgrades, which may crowd out financial resources originally allocated for digital transformation through new infrastructure. Furthermore, increased media attention may raise stakeholders’ demands for short-term performance, leading to myopic decision making and reduced willingness to invest in long-term, high-return NIC projects, ultimately undermining sustainable development outcomes.

5. Heterogeneity Analysis

5.1. Heterogeneity of NIC Sub-Dimensions

Based on the distinct characteristics of different types of NIC, this study categorizes NIC into three dimensions to examine its heterogeneous effects on SDP. The three categories are information infrastructure, integration infrastructure and innovation infrastructure.

As shown in

Table 10. Results of the heterogeneity analysis: infrastructure differences.

Variables	Information Infrastructure			Integration Infrastructure			Innovative Infrastructure
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg	(7) Sdp	(8) Fin	(9) Esg
Nic1	0.0581 *** (4.26)	0.0197 *** (4.21)	0.5388 *** (4.37)
Nic2				0.0665 *** (3.86)	0.0215 *** (3.00)	0.6645 *** (4.46)
Nic3							0.0552 *** (6.60)	0.0096 *** (5.59)	0.5094 *** (8.18)
_Cons	−0.3049 *** (−3.25)	−0.2668 *** (−6.24)	−3.1946 *** (−4.27)	−0.2999 *** (−3.18)	−2.2636 *** (−6.15)	−3.1550 *** (−4.19)	−0.2998 *** (−3.17)	−0.2629 *** (−6.19)	−3.1541 *** (−4.19)
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.3561	0.3910	0.3867	0.3539	0.3599	0.3820	0.3609	0.3977	0.3911

Note: *** represent significant at 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

, the impact of information infrastructure construction (Nic1) on SDP, financial performance (Fin) and ESG performance (Esg) is significantly positive at the 1% significance level, with estimated coefficients of 0.0581, 0.0197 and 0.5388, respectively. Similarly, integration infrastructure (Nic2) on SDP (Sdp), financial performance (Fin) and ESG performance (Esg) was significantly positive at the 1% confidence level, with impact coefficients of 0.0665, 0.0215 and 0.6645, respectively. The regression results of innovation infrastructure (Nic3) are similar, showing a significant positive correlation at the 1% confidence level. The impact coefficients on SDP, financial performance (Fin) and ESG performance (Esg) are 0.0552, 0.0096 and 0.5094, respectively. It can be seen that information infrastructure, integration infrastructure and innovation infrastructure can significantly promote the SDP, financial performance and ESG performance of manufacturing enterprises. Among them, the estimation coefficient of integration infrastructure is the highest in the three sub-dimensions, and the innovation infrastructure is the lowest, which means that integrated infrastructure has the strongest promotion effect on SDP. The promotion effect of innovation infrastructure is the lowest.

The possible reason is that integration infrastructure includes smart energy infrastructure such as ultra-high voltage, smart grid, optical storage charging and discharging integrated system, which can promote the efficiency of clean energy utilization, thus promoting the SDP of enterprises and promote ESG performance. In terms of financial performance, integration infrastructure can strengthen the integration of information networks, empower the real economy through integration, upgrading and synergy of manufacturing enterprises, thus bringing cost savings and income increase in the short term. Innovation infrastructure involves technology research and development, scientific research and other fields. Although it is beneficial in the long run, it has high cost and long return period in the short term. Therefore, its promotion effect on the SDP of enterprises is lagging behind, so the promotion effect is relatively minimal.

5.2. Region Differences

Following the classification by China’s National Bureau of Statistics, the sample is divided into eastern, central, and western regions to explore regional heterogeneity. As presented in

Table 11. Results of the heterogeneity analysis: regional differences.

Variables	Eastern Region			Central Region			Western Region
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg	(7) Sdp	(8) Fin	(9) Esg
Nic1	0.0514 *** (3.80)	0.1078 ** (2.42)	0.4242 *** (5.12)	0.0363 *** (3.76)	0.0161 *** (3.52)	0.2679 *** (3.56)	0.0011 (1.11)	0.0023 (0.84)	0.0957 (1.32)
_Cons	−0.1389 (−0.94)	−0.4218 *** (−5.54)	−0.8525 (−0.77)	−0.4245 (−1.52)	−0.3747 *** (5.86)	−4.1003 * (−1.74)	−0.5282 ** (−2.37)	−0.3024 ** (−2.49)	−6.4798 *** (−3.58)
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.4024	0.4450	0.4417	0.4803	0.5691	0.4815	0.5963	0.4789	0.7109

Note: *, **, *** represent significant at 10%, 5%, 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

, the regression results for enterprises in the eastern and central regions are both positive and largely significant at the level of 1% or 10%, but the coefficients are different. In terms of SDP, the regression coefficients of enterprises in the eastern and central regions are 0.0514 and 0.0363 respectively. In terms of financial performance (Fin), the regression values were 0.1078 and 0.0161, respectively; in terms of ESG performance (Esg), the regression coefficients are 0.4242 and 0.2679, respectively. The above results show that NIC has promoted SDP, financial performance and ESG performance of enterprises in the eastern and central regions, and has a greater role in promoting enterprises in the eastern region. In contrast, the coefficients for the western region are positive but statistically insignificant, indicating that the role of NIC in advancing SDP among manufacturing enterprises in this region remains limited.

This disparity may be attributed to regional developmental imbalances. Enterprises in the eastern region benefit from robust economic foundations, advanced market mechanisms, and preferential policy support, enabling a virtuous cycle between NIC and enterprise SDP improvement. Although central region enterprises possess favorable resource endowments, their capacity for efficient resource allocation is relatively constrained; however, targeted policy interventions have helped drive NIC adoption. In the western region, despite recent policy emphasis on green development, NIC development lags behind overall industrial transformation needs. Infrastructure upgrading has not kept pace, hindering timely innovation and the launch of competitive digital products and services, thereby limiting the effectiveness of NIC in promoting enterprise SDP.

5.3. Enterprise Property Differences

According to the property right of enterprises, this study analyzed the heterogeneous impact of different property right. In terms of enterprise property right classification, the sample is stratified into state-owned enterprise and non-State-owned enterprises.

Table 12. Results of the heterogeneity analysis: enterprise property differences.

Variables	State-Owned Enterprise			Non-State-Owned Enterprises
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg
Nic	0.0217 *** (3.01)	0.0054 *** (3.33)	0.2959 *** (5.82)	0.0394 (0.73)	0.0036 (1.16)	0.0085 (0.21)
_Cons	0.1023 (0.59)	−0.3774 *** (−4.16)	−1.6737 (−1.15)	−0.7091 *** (−4.29)	−0.3352 *** (−5.30)	−7.1137 *** (−5.80)
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.4391	0.2875	0.3252	0.3974	0.4021	0.4296

Note: *** represent significant at 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

shows that in the sample regression of state-owned enterprises, the impact of NIC on SDP, financial performance (Fin) and ESG performance (Esg) is significantly positive at the 1% confidence level, and the impact coefficients are 0.0217, 0.0054 and 0.2959, respectively. It can be seen that NIC can significantly promote the SDP, financial performance and ESG performance of state-owned manufacturing enterprises. In the sample regression of non-State-owned enterprises, the impact of NIC on SDP, financial performance and ESG performance are not significant.

The possible reasons for this situation are as follows: Firstly, the policy dominance of state-owned enterprises. Different from non-State-owned enterprises, state-owned enterprises not only have more development resources, bear more social responsibilities, but can also obtain more policy support. In the long-term strategy, it is more consistent with the high investment of NIC projects. Secondly, state-owned enterprises have stronger anti-risk ability. The new infrastructure technology set updates quickly and has a high risk of failure. State-owned enterprises have policy advantages, stronger financial support and policy tilt, so they are more likely to obtain welfare subsidies and project support. Therefore, in the face of the crisis, state-owned enterprises can use their obvious dominant position to resist the impact of related risks, while non-state-owned enterprises will be subject to more policies, funds and human resources. The speed of digital transformation is relatively slow and the depth is relatively insufficient, so the resilience of the risk of logarithmic intelligent transformation is weak.

5.4. Enterprise Life Cycle Difference

According to the enterprise development stage characteristics of enterprises, this study examines the heterogeneous impact of NIC on enterprises of different life cycle stages. Following standard classification criteria, the sample is categorized into three groups: growth, maturity, and decline phases.

Table 13. Results of the heterogeneity analysis: enterprise life cycle difference.

Variables	Growth Period			Mature Period			Decline Period
	(1) Sdp	(2) Fin	(3) Esg	(4) Sdp	(5) Fin	(6) Esg	(7) Sdp	(8) Fin	(9) Esg
Nic	0.0227 *** (5.18)	0.0082 *** (3.66)	0.2297 *** (6.22)	0.0187 ** (2.40)	0.0051 ** (2.04)	0.2582 ** (2.54)	0.0141 (1.38)	0.0044 (0.71)	0.0585 (0.72)
_Cons	−0.3667 *** (−3.33)	−0.2008 *** (−4.23)	−3.9491 *** (−4.53)	−0.3365 ** (−2.28)	−0.2117 *** (−2.73)	−3.3924 *** (−2.76)	−0.4521 (−1.57)	−0.5753 *** (−3.61)	−4.2460 * (−1.93)
Year Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
Pro × Ind Fe	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes	Yes
R2	0.4578	0.4783	0.4001	0.4070	0.4591	0.4234	0.5029	0.4422	0.5430

Note: *, **, *** represent significant at 1% levels, respectively. Coefficient clustering standard errors (clustered at the province level) are in parentheses.

shows that in the growth phase, the impact of NIC on SDP, financial performance (Fin), and ESG performance (Esg) are significantly positive at the 1% or 10% level, with impact coefficients of 0.0227, 0.0082 and 0.2297, respectively. This indicates that NIC significantly enhances SDP, financial performance and ESG performance of manufacturing enterprises in growth period. At the mature period, the impact of NIC on SDP, financial performance (Fin), and ESG performance (Esg) are significantly positive at a 1% confidence level, with impact coefficients of 0.0187, 0.0051 and 0.2582, respectively. In the decline period, the impact of NIC on SDP, financial performance (Fin) and ESG performance (Esg) are not significant.

The possible reason is that growing enterprises are in the stage of rapid expansion, which can rapidly expand production capacity through NIC and enjoy tax incentives and policy subsidies, so the financial returns are significant in the short term. In addition, growing enterprises tend to pay more attention to innovation and technology investment and have stronger acceptance of new technologies, which makes NIC more effective in promoting their sustainable development. In the mature period, enterprises are limited by path dependence and transformation costs, and new infrastructure transformation and upgrading face greater technical compatibility problems. For example, the data interaction barriers between traditional production lines and industrial internet platforms lead to a smaller improvement in SDP than in the growth period. For enterprises in the decline phase, weakened organizational flexibility, internal governance deficiencies, and shrinking growth opportunities hinder their ability to adapt to technological change. As a result, they struggle to utilize infrastructure upgrades effectively, leading to insufficient momentum for sustainable development and diminished returns from NIC investments.

6. Conclusions and Policy Implications

Based on data from Chinese manufacturing enterprises in 2012–2022, this study empirically tested the impact of NIC on the SDP of manufacturing enterprises. The findings are as follows: (1) NIC significantly enhances the SDP of manufacturing enterprises, financial performance, and ESG performance, a conclusion robust to rigorous tests for robustness and robustness. (2) Mechanism analysis reveals that institutional pressures moderate the relationship between NIC and SDP. Specifically, environmental regulation and market competition strengthen the positive effect of NIC on SDP, whereas media attention attenuates it. (3) Heterogeneity analysis shows that at the infrastructure level, integration infrastructure exerts the strongest promoting effect. At the region level, the positive impact is significant in eastern and central China but statistically insignificant in the western region. At the enterprise level, NIC has a more pronounced effect on state-owned enterprises and firms in the growth stage.

The policy implications of this study are threefold. Firstly, given that NIC serves as a key driver of sustainable development in manufacturing, governments should actively accelerate NIC deployment and emphasize its role in advancing industrial sustainability. Priority should be placed on strategic areas and weak links in infrastructure development, with continuous innovation and integration of core technologies and equipment to address critical “bottleneck” constraints. Second, because institutional pressures influence how NIC affects SDP, policy makers should strengthen internal and external governance mechanisms. Specifically, environmental regulations should be strategically applied to encourage enterprises to adopt ecological practices and implement sustainability strategies. In parallel, incentive-based policies, such as tax incentives, green credit, and sustainable finance, should be introduced to motivate enterprises to proactively deploy NIC for long-term sustainability. Furthermore, the media environment should be optimized to enhance its function as an external governance mechanism, leveraging public opinion to monitor corporate behavior and promote responsible operations. Third, considering the substantial heterogeneity across sectors, regions, and enterprise types, policies should avoid a one-size-fits-all approach. Instead, targeted and differentiated support measures should be designed based on the distinct characteristics of enterprises and regions to maximize policy effectiveness.

This study also has several limitations. First, in measuring NIC, this study adopts an index evaluation method based on the existing literature, which may be relatively coarse. Future research could refine the measurement of NIC by developing more granular and dynamic indicators. Second, the sample is limited to manufacturing enterprises, potentially limiting the generalizability of the findings to small- and medium-sized enterprises (SMEs) or specialized, innovative firms. Future studies should extend the analysis to SMEs and high-tech enterprises to examine whether NIC similarly promotes their SDP and whether differential effects exist compared to listed firms. Third, the operationalization of institutional pressure variables may have limitations. Future work should explore alternative proxy measures and incorporate diverse institutional contexts to deepen understanding of these moderating mechanisms.