The integration of the digital economy and green development makes the coordinated advancement of digital and green transformation ("twin transition") an imperative for manufacturing enterprises. As a new production factor,data is reshaping resource allocation,organizational decision-making,and value creation. Data factor markets,as institutional arrangements for data circulation,may facilitate twin transitions by improving the accessibility,mobility,and utilization efficiency of heterogeneous data resources. However,existing literature primarily focuses on macroeconomic effects of data markets or discusses digital and green transformations separately,leaving insufficient attention to whether and how data factor market development promotes twin transitions and under what boundary conditions such effects become evident.
To address these gaps,this study develops an integrated framework grounded in information ecosystem theory and resource orchestration theory. Information ecosystem theory posits that data circulation is embedded in dynamic systems shaped by institutions,technologies,and multi-actor interactions. Accordingly,data factor market development optimizes the external information environment by reducing transaction frictions,alleviating information asymmetry,and enabling compliant cross-organizational data flows. Resource orchestration theory further explains how firms transform externally acquired data into internal capabilities,emphasizing that data value depends not merely on access but on firms' ability to integrate and leverage such resources across production management,energy utilization,supply chain coordination,and green governance. Accordingly,this study argues that data factor market development can promote the twin transition of manufacturing enterprises by improving the external information ecology and enhancing internal resource allocation efficiency.
Using panel data on Chinese A-share listed manufacturing enterprises from 2010 to 2024,this study takes the staggered establishment of data exchanges across provinces and municipalities as a quasi-natural experiment. To measure enterprises′ twin transition level,it constructs an evaluation index system covering both digital and green dimensions,and applies the entropy weight method together with the coupling coordination degree model. To identify the causal effect of data factor market construction,the paper combines propensity score matching with a multi-period difference-in-differences model,and further conducts parallel-trend tests and placebo tests to ensure the robustness of the empirical results.
The empirical results show that data factor market development significantly promotes the twin transition of manufacturing enterprises. This indicates that the institutionalization of data ownership confirmation,circulation,and trading can effectively activate the synergistic value of data as a production factor,thereby helping enterprises improve digital operating efficiency while embedding green constraints into production and management processes. Mechanism analysis further shows that resource allocation efficiency is an important transmission channel. Specifically,data factor market development enhances enterprises′ ability to integrate external data into internal operations,making production processes more transparent,decision-making more agile,and environmental objectives more measurable and controllable. By contrast,alternative explanations such as technological progress and strategic adjustment do not pass the full mechanism identification test,suggesting that resource allocation efficiency is the more robust and convincing channel.
Further analysis reveals significant heterogeneity in the policy effect. Compared with enterprise-led and mixed models,government-led data factor market development shows a stronger promoting effect. Moreover,the positive effect is more pronounced among non-heavy-polluting firms,non-high-tech firms,and non-state-owned firms. In addition,the promoting effect becomes stronger when firms have higher R&D intensity,when industry competition is more intense,and when the regional level of informatization is higher. These findings indicate that the effectiveness of data factor market development is influenced by enterprise′ absorptive capacity,competitive environment,and regional digital conditions.
This study contributes to the literature in three respects. First,it extends research on data factor market development from the macro level of institutional design and economic efficiency to the micro level of enterprise transformation by providing direct empirical evidence based on a quasi-natural experiment. Second,the study provides an integrated framework explaining how these two dimensions of digital transformation and green transformation mutually reinforce each other under the enabling effect of the data factor market. Third,it clarifies the heterogeneous boundaries of the policy effect across ownership structures,industrial characteristics,and regional environments. Overall,the findings provide useful theoretical support and policy implications for promoting the twin transition of manufacturing enterprises through data factor market construction.

In an era marked by escalating geopolitical volatility,recurrent exogenous shocks like extreme climate events and public health crises,global supply chains face mounting structural vulnerabilities and operational uncertainties. Consequently,bolstering enterprise-level supply chain resilience has evolved from a mere operational concern into a core strategic imperative—central to safeguarding national economic security,stabilizing the real economy and preserving long-term industrial competitiveness in the global market. Concurrently,the rapid advancement of digital transformation has redefined the role of data in economic development:no longer merely a passive byproduct of business operations,data now functions as an active,strategic core production asset that enables agility,real-time responsiveness,and anticipatory decision-making across interconnected supply ecosystems. Within this evolving data ecosystem,government-collected and publicly accessible data plays a pivotal role:it is comprehensive in coverage,inherently non-rivalrous,broadly representative across various sectors and geographies,and underpinned by solid institutional legitimacy that collectively enhances its utility for systemic risk mitigation and industrial collective learning in supply chain management.
Practically speaking,openly available government datasets help fill persistent intelligence gaps in key domains such as scientific demand forecasting,multi-dimensional supplier due diligence,and pre-emptive risk identification and early warning. By integrating these authoritative public resources into daily operational decisions,firms can optimize their inventory holding policies,dynamically recalibrate sourcing and sales networks in near real time,and orchestrate more agile,evidence-based responses during unexpected supply chain disruptions,thereby significantly strengthening both organizational adaptability and broader supply network recovery capabilities. As such,understanding the causal relationship between governmental data transparency and corporate supply chain resilience,and clarifying its internal transmission mechanisms,is not only academically compelling but also critically timely for formulating evidence-informed policymaking to boost industrial chain security.
Drawing on a balanced panel dataset of Chinese A-share listed companies spanning 2007–2024,this study takes the launch of local government data open platforms as a quasi-natural experiment,and employs fixed-effects,instrumental-variable approaches and a series of rigorous robustness tests to causally identify the impact of government data opening on enterprise supply chain resilience. Results indicate that government data opening significantly and robustly strengthens enterprise supply chain resilience,with precise supply-demand matching emerging as a key and core mediating mechanism that bridges data openness and resilience enhancement. Moderation analyses further reveal that both enhanced digital infrastructure and a mature and sound legal framework for data significantly intensify this positive effect by providing technical support and institutional guarantee respectively. The heterogeneity analysis indicates that in the groups of small and medium-sized enterprises,private enterprises,enterprises with weak market positions,and those in regions with low marketization levels,the enhancing effect of government data opening on the resilience of enterprise supply chains is more significant and prominent,as these entities suffer more from information asymmetry and resource constraints. The expansion analysis shows that the improvement of the overall quality of government data opening,especially in the critical subdivision dimensions of policy guarantee,data quality,and platform construction,all independently and significantly strengthen the resilience of enterprise supply chains,verifying the vital importance of a quality-oriented development model in government data opening practices.
This study proposes four targeted and actionable policy recommendations for practical implementation. Governments should first accelerate the construction of government data opening platforms and unify national construction standards to fully activate the value of public data elements. Second, authorities need to strengthen the digital infrastructure and institutional guarantees for precise supply-demand matching, and vigorously promote the construction of new digital infrastructure and the improvement of the data legal system.Third, a targeted data capability support system ought to be established to enhance the data utilization efficiency of information disadvantaged entities such as SMEs and private enterprises. Fourth, governments shall keep improving the quality of government data opening and complete relevant supporting mechanisms, comprehensively boosting the reliability, usability and practical application value of publicly available government data.

As the core carrier of national strategic scientific and technological strength, technology-leading enterprises serve as the main force and leader in breaking through the "bottleneck" of key core technologies and achieving the alternative innovation of key core technologies. Alternative innovation theory provides new theoretical underpinnings for systematically deconstructing the process of domestic substitution from technological breakthroughs to market application. Compared to ordinary enterprises, technology-leading enterprises possess unique advantages such as powerful resource integration capabilities and globally leading innovation R&D systems. Consequently, a deep investigation into the intrinsic mechanism underpinning their leadership in alternative innovation of key core technologies holds significant theoretical and practical implications.
From the theoretical perspective of multiple institutional logics and using a case study approach, this research focuses on the practice of alternative innovation in international mobile communication standards led by China Mobile Group, to systematically analyze the mechanisms of alternative innovation for key core technologies led by technology-leading enterprises. The access channels for primary and secondary data are well-established. Primary data were collected from November 2024 to May 2025 through semi-structured in-depth interviews with seven core relevant personnel from China Mobile Group and internal documents, accumulating 390 minutes of interviews and generating 122 000 words of transcribed text. Secondary data encompassed corporate annual reports, official information, media coverage, and industry publications, totaling approximately 525 000 words. A complete chain of evidence was formed through cross-validation between primary interview data and secondary public sources. Data analysis employed the analytical induction method for case study research, integrating theories of substitution-based innovation in key core technologies and multiple institutional logics, sequentially conducting first-order initial coding, second-order thematic categorization, and core theoretical dimension refinement. Simultaneously, research reliability was strengthened through double-blind coding, anonymization processing, and third-party expert evaluation of divergence points, ensuring deep connections among data, phenomena, and theory, as well as transparent theoretical construction processes and reliable research conclusions.
The findings reveal that, first, the alternative innovation of key core technologies led by technology-leading enterprises demonstrates characteristics of the new nationwide system, where the state plays a strategic guiding role, research institutions undertake the task of technological breakthroughs in the initial phase, and technology-leading enterprises play a central role in technological upgrading and market substitution. Second, technology-leading enterprises achieve the upgrading and market substitution of key core technologies by aligning with state logic and integrating technological logic with market logic. Third, during the process of alternative innovation for key core technologies, multiple institutional logics demonstrate characteristics of dynamic co-evolution and mutual reinforcement.
The theoretical contributions of this paper are as follows: First, it constructs a theoretical model of alternative innovation for key core technologies led by technology-leading enterprises. Second, this research deepens and enriches the theory of alternative innovation while enriching the literature on technology-leading firms and key core technologies. Third, it reveals the mechanisms of dynamic coordination and mutual reinforcement among multiple institutional logics in alternative innovation processes, thereby advancing existing research on multiple institutional logics.
To achieve alternative innovation in key core technologies, the practical recommendations are as follows: Government departments should rely on the new nationwide system to refine top-level strategic planning and provide dynamic resource support. Technology-leading enterprises should leverage their core capabilities and endowments to play a leading role in driving innovation of key core technologies. Research institutes, in turn, should focus on national strategic needs, undertake technological breakthrough tasks in the initial phase, and maintain close collaboration with technology-leading enterprises. This study not only expands the explanatory scope of the multiple institutional logics theory, but also clarifies the alternative innovation mechanism for key core technologies led by technology-leading enterprises.

In the context of global scenario-driven transformation of technological competition, identifying high-potential scenarios that can integrate technological innovation, industrial development, and policy orientation has become a key challenge for emerging industries. Traditional scenario studies exhibit two major limitations: single-agent perspective and static spatiotemporal analysis, which constrain understanding of technological evolution and weaken guidance for innovation practice. This study develops a scenario mining framework based on multi-subject conceptualization, integrating perspectives of technology developers, industry practitioners, policymakers, and end-users to systematically connect scenario-driven needs with technological evolution trajectories and support innovation ecosystem construction.
In terms of methodology, this study combines text mining, patent analysis, and network modelling to ensure both theoretical rigor and empirical validity: (1) Within the multi-subject conceptual framework, multi-source heterogeneous textual data (including policy documents, industrial white papers, scientific reports, and user feedback) are analyzed. The technology-relationship-technology (TRT) semantic analysis method is adopted, which identifies scenario elements embedded in linguistic structures by focusing on prepositional relationships that define the expression of specific scenarios in context. After repeated semantic extraction, a list of high-potential scenarios is systematically constructed; (2) The KeyBERT algorithm is used to extract technical terms from the Novelty section of patents retrieved from the Derwent Innovation Index, establishing a scenario-technology correlation map that links high-potential scenario requirements to their underlying technological support; (3) The study constructs a "technology network-evolutionary cycle" analytical model, integrating social network analysis and technology life cycle theory. This model captures the spatiotemporal evolution of technology networks under the influence of high-potential scenarios, providing a dynamic perspective on how technologies co-evolve and mature within the innovation ecosystem.
The empirical analysis focuses on the field of intelligent driving, a highly dynamic and data-rich area of emerging technology convergence. By drawing upon 24 378 active patents and multiple textual data sources, this study identifies eight categories and 53 sub-scenarios of intelligent driving applications, covering extreme environment sensing, human-computer interaction, collaborative vehicle-road systems, AI-based decision-making, and system resilience and safety. The empirical results yield several key findings: (1) In the spatial dimension, the technology network exhibits prominent scenario-dependent differences. Core technologies, such as autonomous driving algorithms, sensor fusion, and environmental perception form dense central clusters, while peripheral technologies (e.g., vehicle-road coordination and urban traffic management) are more dispersed. Scenarios such as intelligent traffic management and collaborative driving demonstrate strong technological synergies, whereas severe weather sensing scenarios rely more on advances in sensing and data processing capabilities. This spatial differentiation highlights the multi-level structure and collaborative dynamics of technological innovation across scenarios. (2) In the temporal dimension, the development of intelligent driving technology is closely coupled with the S-shaped technology life cycle. In the early stage, technology develops slowly, focusing primarily on algorithmic fundamentals and system framework construction. With the advancement of artificial intelligence, machine learning, and perception algorithms, the field of intelligent driving has entered a stage of rapid growth. These developments show that core technologies are proliferating and maturing rapidly, confirming that the technology development pattern follows a typical S-curve trajectory.
This study provides a comprehensive analytical framework connecting multi-agent scenario cognition with technology network dynamics and lifecycle evolution.The results deepen theoretical understanding of how high-potential scenarios drive technology evolution while providing actionable guidance for promoting collaborative innovation ecosystems and sustainable industrial transformation. At the theoretical level, this study expands the paradigm of scenario-driven technological innovation analysis by extending scenario research from static single-subject models to dynamic multi-subject multi-source analytical frameworks. At the practical level, the results provide decision support for policymakers and industry leaders, helping identify key technologies, rationally allocate innovation resources, and design forward-looking innovation policies.

The rapid growth of AI and the digital economy has made computing power a strategic resource for industrial transformation. China's intelligent computing capacity reached 725 EFLOPS in 2024,yet resources remain unevenly distributed,with eastern cities dominating the top tier. Computing-power technology innovation is emerging as a core pathway to break these constraints,enabling large-scale and more efficient application of advanced technologies while reshaping inter-industry collaborative development patterns. Unlike conventional digital innovation that clusters in coastal hubs,computing-power innovation is factor-driven and relies on energy,climate,and land endowments,while its geographically transferable loads enable cross-regional transmission. Existing research on computing-power technology innovation has predominantly examined its macroeconomic impacts and enterprise-level applications,leaving the micro-level transmission mechanisms that drive coordinated regional development largely unexplored. This raises critical questions:how does computing-power technology innovation reshape coordinated regional development through spatial spillovers? Through what mechanisms does this influence operate across different geographic scales?
Drawing on spatial economic theory and growth-pole analysis,this study examines the effects and mechanisms of computing-power technology innovation on coordinated regional development in China. The study uses panel data from 282 prefecture-level cities (2005–2021),and constructs a city-level computing-power technology innovation indicator by expanding a 36-term industry-standard lexicon with BERT and identifying 47 104 computing-power patents through an LLM-based classifier,which are then geocoded and aggregated to the prefectural level. Coordinated regional development is measured at multiple scales using 500-meter VIIRS-calibrated nighttime light data and 1-kilometer gridded population data. The empirical analysis applies single-regime and two-regime spatial Durbin models with time and city-fixed effects,complemented by threshold distance matrices to trace the nonlinear decay of spillovers. The findings indicate that computing power technology innovation produces significant positive spatial spillovers,raising the level of coordinated regional development and narrowing the relative development gap in administrative boundary areas,with marginal effects on boundary townships notably exceeding those on central townships. Regarding mechanisms,inter-industry co-agglomeration and the release of human capital agglomeration dividends serve as important paths,reshaping the spatial distribution into a polycentric and specialized pattern. Heterogeneity analyses show that the coordinating effect strengthens in cities with open public data platforms,whereas high entrepreneurial activity amplifies the siphoning effect. Further analysis shows that the spillovers of computing power innovation growth poles follow nonlinear geographic decay:within a 145-kilometer threshold,growth poles significantly promote neighboring cities' coordinated development,whereas beyond 160 kilometers the backwash effect dominates.
The contributions of this paper are reflected in three aspects. First,while most existing studies on digital economy and regional development focus on generic digital technologies or the macro-impacts of digital infrastructure,few studies define and measure computing power technology innovation as a distinct factor-driven paradigm. This study clarifies its connotation,uses LLM techniques to identify computing-power patents from large-scale textual data,and measures city-level computing power technology innovation at fine spatial resolution,thereby providing a new analytical construct and empirical basis for research on computing-power economics. Second,existing research rarely combines multi-scale spatial analysis with the geographically transferable properties of computing power. This study addresses the gap by combining spatial Durbin models with nighttime light data and gridded population data,jointly capturing inter-city spillovers and intra-city center and periphery dynamics at prefectural,township,and urban-core scales,thereby opening the black box of how computing power technology innovation affects coordinated regional development through inter-industry co-agglomeration and human capital redistribution. Third,existing research on regional growth poles mainly focuses on conventional innovation hubs,and few studies examine how computing power innovation growth poles shape surrounding cities under different geographic-distance conditions. Aligned with China's strategic orientation toward national supercomputing centers,this study builds a research framework for computing power innovation growth poles and empirically identifies the nonlinear distance-decay structure of their spillovers,providing spatial-planning guidance for national computing infrastructure and evidence-based support for reducing regional development disparities.

Against the backdrop of global technological revolution and industrial transformation, the smart economy driven by artificial intelligence as its core and representing an advanced form of the digital economy has emerged as a central engine for China′s high-quality economic development and the construction of a modernized economic system. To promote the deep integration of AI and the real economy, China has implemented the strategic initiative of National New Generation Artificial Intelligence Innovation and Development Pilot Zones (hereinafter referred to as “AI Pilot Zones”). Existing research has rarely directly examined the causal relationship between this pilot policy and the development of the smart economy, nor has it systematically explored the underlying transmission mechanisms, spatial spillover effects, and heterogeneous impacts. Thus, this study evaluates the enabling role of AI Pilot Zones in the smart economy, explores their internal mechanisms and spatial dynamics, and offers empirical insights to optimize AI policy frameworks and promote regional coordinated development.
Utilizing provincial-level panel data from 30 provinces (municipalities and autonomous regions) in China spanning the period from 2011 to 2023, this study constructs a comprehensive evaluation index system comprising four dimensions:smart resource foundation, smart factor integration, smart industry application, and smart environment support. The entropy-weighted TOPSIS method is employed to measure the smart economy development level of each province. To identify the net effect of the pilot policy on smart economy development, this study treats the establishment of AI Pilot Zones as a quasi-natural experiment and adopts a time-varying difference-in-differences model as the baseline regression method. A mediation model is further used to examine the underlying mechanisms, and a spatial Durbin difference-in-differences model is applied to investigate spatial spillover effects. To ensure the robustness of the empirical findings, multiple endogeneity tests and robustness checks are conducted, alongside heterogeneity analyses from the perspectives of regional location and the degree of marketization.
The empirical results yield the following findings: First, the AI Pilot Zone policy significantly promotes the development of the regional smart economy, a conclusion that remains robust after a series of endogeneity and robustness checks. Moreover, the AI Pilot Zones significantly enhance the development levels of the four sub-dimensions of the smart economy:smart resource foundation, smart factor integration,smart industry application, and smart environment support. Second, the policy exerts its impact through three key mediating channels: enhancing human capital levels, fostering digital industry agglomeration, and mitigating the misallocation of innovation factors. Specifically, the pilot policy attracts high-end talent and improves labor quality, promotes the agglomeration of digital industries leading to economies of scale and knowledge spillovers, and optimizes the allocative efficiency of innovation resources, thereby jointly driving the growth of the smart economy. Third, China′s smart economy development exhibits significant positive spatial autocorrelation. The AI Pilot Zone policy generates notable positive spatial spillover effects, not only enhancing the smart economy level of the host locality but also effectively stimulating the development of smart economies in neighboring regions. Fourth, significant heterogeneity characterizes the policy effects: the promotional effect is strongest in the Western region, followed by the Eastern, Central, and Northeastern regions in a descending order. The policy dividend is more pronounced in areas with higher degrees of marketization, while the effect is relatively weaker in regions with lower marketization levels.
The study links the AI Pilot Zone policy to the smart economy through a quasi-natural experiment, clarifying mechanisms and spatial spillover effects while providing a scientific basis for differentiated policymaking and coordinated regional advancement. The study recommends expanding AI Pilot Zones with resource tilted toward the Central, Western, and Northeastern regions, strengthening talent-industry linkages to optimize innovation allocation, and leveraging spatial radiation effects alongside market-oriented reforms to sustain AI's long-term driving force on the smart economy.

In the digital economy era, propelled by the dual forces of technological evolution and policy guidance, digitally-enabled collaborative innovation has become a pivotal strategic imperative for enterprises to cultivate competitive advantages and achieve sustainable development.Digital technology diversification empowers firms to enhance inter-organizational resource orchestration and opportunity responsiveness through constructing digital technology portfolios, expanding digital-physical collaborative frontiers, and dynamically allocating heterogeneous technological elements, thereby serving as a critical value co-creation source of collaborative innovation in digital context.However, current research has three gaps:first, it ignores the actual organizational behaviors of deploying multiple digital technologies simultaneously; second, it remains unclear that whether digital technology diversification can mitigate the limitations of conventional technological diversification through complementary and synergistic effects of technologies; third, it lacks a systematic analysis of how internal organizational capabilities(absorptive capacity) and external environmental conditions (environmental munificence) contingently shape the enabling potential of digital technology diversification for collaborative innovation.
Following technology affordance theory, this study conducts empirical analysis using a sample of A-share listed high-tech firms in China from 2012 to 2023.The study identifies sample firms based on high-tech enterprise certification information disclosed in the CSMAR database, following a rigorous screening process,which yields an unbalanced panel of 3 211 high-tech enterprises comprising 21 059 firm-year observations.Patent data for measuring digital technology diversification and collaborative innovation are sourced from Clarivate Analytics' IncoPat database, while firm-level and industry-level data for absorptive capacity, market munificence, and control variables are obtained from CSMAR.To address unobserved firm-specific factors (e.g., corporate culture) and time-varying factors (e.g., economic cycles), mitigate potential endogeneity, and enhance estimation accuracy, the study employs two-way fixed effects models with both firm and year fixed effects for hypothesis testing.
This study finds that digital technology diversification significantly enhances firms' collaborative innovation.The relationship is positively moderated by absorptive capacity, meaning the effect is stronger when firms possess higher capacity to absorb knowledge.Conversely, market munificence negatively moderates the relationship, indicating that the positive impact is weaker in more favorable market environments.These moderating effects are evidenced by significant interaction terms in regression models.To ensure robustness, the study conducts multiple supplementary tests.These include re-measuring digital technology diversification using all patent IPC codes, extending the lag structure to two years, and operationalizing collaborative innovation via the natural logarithm of jointly granted patents.The results across these various specifications consistently support the baseline conclusions regarding the effects of diversification, absorptive capacity, and market munificence.
The theoretical contributions of the study are as follows:First, this study extends the research of technological diversification into the digital context by clarifying the concept of digital technology diversification and validating the applicability of technology affordance theory within this domain.Second, this study shifts the research focus from digital technologies as an aggregate or single-dimensional construct to diversified digital technology portfolios and reveals the positive impact of digital technology diversification on collaborative innovation, thereby deepening the understanding of collaborative innovation in digital contexts.Third, this study elucidates how absorptive capacity and market munificence shape the value realization of collaborative innovation empowered by digital technology diversification, advancing the comprehension of the interactive mechanisms among technology, capability and environment.Besides, the findings of this study also provide practical guidance for firms to optimize technology allocation and enhance collaborative innovation capabilities in the context of digital economy.For managers, it is essential to establish a standardized, modular digital technology framework.Strengthening organizational absorptive capacity is also crucial.This can be achieved by forming cross-functional teams, promoting internal knowledge sharing, and building a knowledge base to enhance the efficiency of integrating diverse technologies.Furthermore, strategies must adapt to market conditions.In resource-rich markets, companies should focus R&D on core technologies and deepen collaboration with key partners.Conversely, in resource-constrained environments, firms should increase the diversity of their digital technologies to broaden the scope for collaborative innovation and improve matching capabilities with external partners.

In an era marked by intensified global economic volatility, a complex and ever-changing trade environment, and heightened geopolitical risks, establishing a self-reliant and controllable industrial and supply chain has become a strategic priority for promoting high-quality development. Supply chain spillover effects have long been recognized as a crucial force driving corporate innovation. Nevertheless, whether artificial intelligence technology, characterized by its unique synergy and generative nature, can produce such spillover effects, as well as the differentiated mechanisms through which it exerts impacts on upstream suppliers and downstream customers, remains insufficiently explored in existing research. To fill this research gap, this study aims to investigate whether and through which pathways artificial intelligence (AI) technology innovation by focal enterprises can act as an external driver for supply chain partners to conduct AI technology innovation activities, so as to provide theoretical support and practical guidance for supply chain collaborative innovation.
Based on the theoretical analytical framework of "supplier-focal enterprise-customer", this study takes China's A-share listed manufacturing companies from 2007 to 2024 as research samples, and conducts an empirical analysis combined with manually compiled artificial intelligence patent data. Finally, 1 316 supplier-focal firm observations and 1 584 customer-focus firm observations are obtained. In the process of data processing and indicator construction, machine learning methods are employed to systematically develop indicators of AI technology innovation. First, AI-related patents are screened from the Incopat patent database using patent IPC codes and keyword matching, covering core technical fields such as machine learning, natural language processing, computer vision, and intelligent robots. Second, semantic features are extracted from patent texts via the BERT pre-trained language model, and a firm-level overall index of AI technology innovation is constructed by integrating patent citations, the number of patent claims, technological scope and other dimensions. On this basis, a technological disruption identification algorithm is further adopted to classify AI technology innovation into two categories:disruptive innovation and incremental innovation. Specifically, disruptive innovation is defined as dual breakthroughs in technology and market, meaning that patented technologies achieve a significant technological leap on the basis of existing knowledge and can open up new market application scenarios. Incremental innovation is characterized by the optimization and upgrading of existing technologies, focusing on performance improvement, scenario adaptation and process transformation of established AI technologies, so as to systematically compare the heterogeneous spillover pathways and mechanisms of the two types of innovation.
The empirical results show that AI technology innovation by focal enterprises exerts a significant positive spillover effect on both upstream and downstream enterprises in the supply chain, and this effect is mainly transmitted through incremental innovation, while the spillover effect of disruptive innovation is not verified. Mechanism analysis reveals heterogeneous transmission channels between the two. For upstream suppliers, AI technology innovation by focal enterprises forces them to increase AI investment, promote industry-university-research cooperation and optimize the financing ecosystem, forming a "forcing effect". For downstream customers, AI technology innovation by focal enterprises stimulates their AI investment, optimizes human resource structure and enhances informatization levels, generating a transmission effect. Heterogeneity tests further indicate that this spillover effect is more pronounced among small-scale suppliers and suppliers located in regions with lower government intervention.
This study makes three main contributions. First, it expands the research dimensions and analytical framework of supply chain spillover effects, extending the research on the impacts of AI technology innovation from the single-firm level to inter-organizational dynamic relationships, thus enriching the theoretical perspective of supply chain collaborative innovation. Second, it refines the classification of AI technology innovation types and constructs targeted measurement indicators, addressing the inadequacy of existing literature in examining the details of firm-level behavioral spillovers. Third, it reveals the differentiated spillover mechanisms of focal enterprises on upstream and downstream partners, providing theoretical support for enterprises to formulate precise cooperation strategies and promote supply chain collaborative innovation.

Against the backdrop of continued global climate governance and deepening consensus on the carbon peaking and carbon neutrality("dual carbon") goals, green transformation is a key path for enterprises to break through resource and environmental constraints and achieve high-quality economic development. As the core subject of resource consumption and pollution emissions in the industrial system, heavily polluting enterprises have long relied on the traditional development model of high investment and high emissions, which not only exacerbates the pressure on the ecological environment, but also traps enterprises in the development dilemma of rising resource costs and weakened market competitiveness. Currently, the green transformation of heavily polluting enterprises faces multiple challenges, including insufficient institutional guidance, weak industrial collaboration, and a lack of endogenous motivation. These issues have led to prominent problems such as a lack of transformation drive and misalignment with viable transformation pathways. Therefore, breaking through this development deadlock and enhancing the efficacy of green transformation has become an urgent issue.〖HJ*3〗
　　Green technology innovation serves as a key driver for corporate green transformation, significantly reducing resource consumption and pollution emissions while promoting industrial upgrading and fostering synergistic development of environmental and economic benefits. Existing literature primarily focuses on single-element analysis, either examining institutional environments or investigating the impact of corporate resource endowments on green technology innovation, overlooking the interactive effects among institutions, industries, and enterprises. It remains unclear whether any individual factor constitutes a necessary condition for achieving high levels of green technology innovation. How do the IIE(short for Institution, Industry, and Enterprise) factors couple to foster high green technology innovation What are the mediating mechanisms of green technology innovation Addressing these questions helps managers formulate differentiated green technology innovation and transformation strategies based on the institutional and industrial environments of their regions, combined with the unique characteristics of their enterprises.
Drawing on complex systems theory and the "strategic triangle" framework, this study analyzes 151 A-share listed companies in China's heavily polluting industries. Using a complex mediation to explore the driving mechanisms of green technology innovation and enterprise green transformation under the three-dimensional interaction of the "institution-industry-enterprise" framework, the study yields the following findings:(1)No single factor constitutes a necessary condition for achieving high green technology innovation; however, enterprise nature exerts a pervasive influence across various contexts.There are three differentiated driving configuration paths for high green technology innovation in heavily polluting enterprises, namely:industry-led enterprise-response type, industry-enterprise dual-drive type, and institutional response type under industry-enterprise dual. There are three paths for non-high green technology innovation, all of which include non-high enterprise nature; (2) The complex mediation mechanism test shows that all six configuration paths that generate high and non-high green technology innovation can indirectly promote the green transformation of heavy polluting enterprises through green technology innovation. The industry enterprise dual wheel drive configuration path in the high configuration has a direct promotion mechanism for the green transformation of enterprises, while the three types of configurations that generate non-high green technology innovation have a direct inhibition mechanism for the green transformation.
The innovation of this article is as follows:(1) Traditional research often explores the influencing factors of green technology innovation from a single perspective of institutions or resources. This study adopts the "strategic triangle" theory to examine multiple driving paths from an "institution-industry-enterprise" configurational perspective, enriching the research on the influencing factors of green technology innovation in heavily polluting enterprises. (2) Existing literature mostly focuses on the linear mediation mechanism of green technology innovation, and there is a lack of research on complex systems formed by multidimensional and interdependent independent variables. This article constructs a complex mediation model to reveal the complex mechanism of how the IIE ecosystem affects the green transformation of heavily polluting enterprises through green technology innovation within the "strategic triangle" framework. It deepens the contextual application of the complex mediation model and expands the research on the driving mechanism of enterprise green transformation.

Promoting the deep integration of the real and digital economies,cultivating advanced manufacturing clusters,and driving the high-end,intelligent development of manufacturing constitute the essential pathway for China to solidify its real economy foundation,build a modern industrial system,and achieve new industrialization.However,amid the digital economy surge,digital transformation has presented manufacturing firms with unprecedented challenges,and many firms find themselves in a predicament where they are reluctant to initiate transformation due to fear of failure or insufficient return on investment.In order to break through this impasse and ensure sustainable,healthy enterprise development,CEOs with information technology backgrounds are assuming increasingly prominent roles.Possessing the dual identity of "executive plus technologist" and wielding digital leadership,they can steer enterprise digital transformation and provide specific technical guidance to dismantle the IT barriers that manufacturing firms may encounter in the digital evolution.
Adopting a human capital perspective,this study investigates how CEOs with IT backgrounds affect manufacturing firm performance.Grounded in upper echelons theory,digital innovation diffusion theory,and digital leadership research,the study explores how these executives leverage digital strategic thinking,digital asset operation capabilities,and digital talent cultivation to drive digital innovation,overcome the "IT paradox", and ultimately enhance organizational performance.First,drawing upon theoretical and empirical literature,this paper constructs a theoretical model in which CEOs′ IT background influences manufacturing enterprise performance,with enterprise digital innovation as the mediating variable and the degree of digital transformation as the moderating variable.Second,utilizing a sample of A-share listed manufacturing firms in Shanghai and Shenzhen from 2015 to 2023,this study empirically investigates the impact and underlying mechanisms of CEOs′ IT background on firm performance,while further exploring heterogeneous effects across ownership structures and industries.Data regarding firm performance,CEOs′ IT background,and digital innovation were collected from the CSMAR and CNRDS databases,as well as annual reports.Following standard screening procedures,the final sample comprises 15 316 firm-year observations across 2 942 enterprises,including 1416 observations featuring CEOs with an IT background.
Through empirical tests,this study obtains the following conclusions: (1) CEOs′ IT background has a significant positive impact on manufacturing firm performance; (2) digital innovation plays a partially mediating role between CEOs′IT background and manufacturing firm performance; (3) the degree of digital transformation plays a positive moderating role in the relationship between CEOs′IT background and corporate digital innovation; (4) in state-owned firms,and high-tech industries,the mediating role of CEOs′IT background on corporate digital innovation as well as digital innovation varies at the nature of property rights,which manifests itself as better performance in non-state-owned firms.
On the basis of these findings,this study offers the following implications: First,as core drivers of digital transformation,CEOs should cultivate digital leadership and develop "IT+X" interdisciplinary competencies,while leveraging such leadership to attract digital talent and build high-caliber innovation teams.Second,firms,particularly those in high-tech industries and state-owned enterprises,should recognize the value of IT-background CEOs and optimize executive teams′ digital capabilities through external recruitment and internal development; they should also strengthen digital infrastructure and innovation investment to create enabling conditions for CEO leadership.Third,the government should foster "IT+X" talent through curriculum optimization and industry-academia collaboration,while providing tiered policy support for high-tech manufacturing innovation and traditional firm digital transformation.
This study integrates relevant research and theories on digital innovation diffusion to discuss the mechanisms through which CEOs with IT backgrounds affect corporate digital innovation and performance,emphasizing the pivotal role of leadership in corporate development.Furthermore,this study also expands the antecedent variables of enterprise performance in the era of digital economy.The findings provide theoretical guidance for firms in selecting CEOs,focusing on digital talent cultivation,and accelerating the improvement of their digital talent systems.

Emerging technologies such as artificial intelligence are reshaping the operating models of various organizations at an unprecedented pace, and their application has become a key driver of organizational innovation and development. In the age of artificial intelligence, intrapreneurship among employees is of great strategic significance for enhancing organizational competitive advantages and facilitating organizational digital transformation. AI disruption awareness refers to employees' perception of threats and concerns when facing AI technology in a digitally intelligent workplace, and their ability to practically apply it has become a key indicator for measuring the effectiveness of a company's intelligent transformation. When enterprises require employees to utilize AI technology for intrapreneurship, AI disruption awareness serves as an important driving force. It is evident that AI disruption awareness is a crucial foundation for employees to engage in intrapreneurship, and exploring its conversion mechanisms is of significant importance. Existing research indicates that AI disruption awareness has a bidirectional impact: negative perceptions lead to negative behaviors, while positive aspects can enhance innovation and learning capabilities. However, there are gaps in the understanding of the influence mechanisms, with few studies examining its impact on intrapreneurship behavior. Employee intrapreneurship is an active transformative behavior, so the creative role identity is introduced as a mediating variable. Based on the person-environment fit theory, job variety and job autonomy as situational factors influence the relationship, and this study incorporates them into the research.
This empirical study employed a two-stage data collection approach. The first stage involved surveying active personnel within government institutions in southwestern China through a combination of electronic and paper questionnaires. The second stage supplemented the sample by recruiting corporate employees via an online platform. Across both stages, 588 questionnaires were distributed, with 417 valid responses ultimately retrieved, yielding a response rate of 70.92%. At the data processing stage, validation factor analysis, common method bias test, descriptive statistical analysis and hypothesis testing were conducted using AMOS26.0 and SPSS27.0 software.
This paper concludes the following:(1) there is a significant positive impact of AI disruption awareness on employee intrapreneurship.(2) Creative role identity plays a mediating role between AI disruption awareness and employee intrapreneurship.(3) job variety has a positive moderating effect on the relationship between AI disruption awareness and creative role identity; the higher the level of job variety, the more pronounced is the promotion of AI disruption awareness on creative role identity, and job variety moderates the mediating effect of creative role identity between AI disruption awareness and employee intrapreneurship.(4) job autonomy also positively moderates the relationship between AI disruption awareness and creative role identity, and the higher the level of job autonomy, the more significant the promotion effect of AI disruption awareness on creative role identity. At the same time, job autonomy also moderates the mediating effect of creative role identity between AI disruption awareness and employee intrapreneurship.
In summary, this study makes the following contributions: First, it extends intrapreneurship research from traditional enterprises to the public sector, exploring whether AI disruption awareness can enhance employees intrapreneurship behavior, thereby expanding the application scenarios and boundaries of intrapreneurship theory. Second, it shifts the research focus from the general positive effects of AI disruption awareness to its underlying mechanisms driving intrapreneurship behavior, introducing creative role identity as a mediating variable to reveal the underlying connections, thereby enriching the theoretical explanation of employee intrapreneurship behavior in the digital age. Third, drawing on the person-environment fit theory, the study introduces the key contextual element of work characteristics in the digital age, verifying the moderating role of job variety and job autonomy in the influence of AI disruption awareness on employees' Intrapreneurship processes. This expands the theoretical understanding of the role of context in the relationship between AI disruption awareness and individual behavior.

As corporate ESG performance evolves from a peripheral sustainability agenda to a core value reconfiguration mechanism, firms increasingly grapple with a paradox where heightened disclosure coexists with pervasive ESG greenwashing, eroding stakeholder trust and market integrity. The 2025 Central Economic Work Conference stressed fostering new quality productive forces and comprehensive green transformation, yet many enterprises remain trapped in prioritizing form over substance, favoring cosmetic reporting over strategic ESG integration. Prior research extensively investigates external institutional drivers and board characteristics that shape ESG responsibility fulfillment or greenwashing separately, but it overlooks the cognitive foundation at the organizational apex—executives' green perception, embedded beliefs, values, and knowledge regarding ecological sustainability and social responsibility. Crucially, how this strategic cognition simultaneously shapes genuine ESG responsibility and cosmetic greenwashing, and the mechanisms through which such cognitive orientations translate into firm conduct, remain underexplored. Integrating upper echelons theory and strategic cognition theory, this study proposes a dual-pathway framework of promoting genuine ESG responsibility while curbing greenwashing, hypothesizing that executives' green perception enhances corporate ESG performance by fostering substantive responsibility fulfillment while concurrently inhibiting greenwashing. Using a sample of Chinese A-share listed firms from 2012 to 2023, the study further examines the mediating role of corporate integrity culture, grounded in social norm theory, and the contingent effects of ownership structure and industry pollution attributes.
Findings demonstrate that executives' green perception significantly improves overall ESG performance through two mutually reinforcing mechanisms:it fosters proactive ESG responsibility fulfillment and effectively suppresses greenwashing behaviors.Endogeneity concerns are addressed using propensity score matching techniques, and a battery of robustness checks confirms that the findings remain consistent. Heterogeneity analyses reveal that the beneficial impact is more pronounced in non-state-owned enterprises and non-heavily polluting industries. In state-owned enterprises, rigid bureaucratic mandates and multi-layered accountability systems constrain the discretionary enactment of green cognition, whereas in heavily polluting sectors, intense regulatory pressure and short-term profit incentives dilute cognitive-driven authentic ESG improvements. Mechanism tests confirm that corporate integrity culture partially mediates both the positive link with ESG responsibility fulfillment and the negative link with ESG greenwashing. It channels executives' green perception into higher responsibility by instilling ethical norms and long-term stewardship values, while simultaneously strengthening organizational commitments to truthful disclosure and substantive action, thereby reinforcing the suppression of greenwashing.
Three implications for managerial practice and policy refinement emerge. First, firms should institutionalize the deep cultivation of executives' green cognition by incorporating cognitive psychology and greenwashing case deconstruction into leadership development programs, and by establishing asymmetric incentive schemes where verified absence of greenwashing serves as a precondition for equity-based compensation and bonuses, thereby converting cognitive commitment into genuine ESG authenticity. Second, given the attenuated effects in state-owned and heavily polluting contexts, differentiated strategies are necessary. Non-state firms should capitalize on their strategic flexibility to embed green cognition into core strategies, while heavily polluting firms must implement rigorous internal ESG audit mechanisms and redirect resources from cosmetic green marketing toward substantive clean production upgrades, with support from regulatory sandboxes and targeted green subsidies. Third, integrity culture must be elevated from a peripheral corporate value to an organizational cornerstone. Enterprises should establish independent board-level ESG oversight committees and embed green cognition into codes of conduct and performance evaluation systems, thus transforming ESG from a compliance cost into a value generation engine and ensuring the sustained authenticity of ESG performance.
This study offers three primary theoretical contributions. First, it enriches upper echelons and strategic cognition perspectives by constructing a dual-dimension framework of promoting ESG substance and curbing greenwashing, while advancing theory on the enabling and constraining roles of executive green perception. Second, it further identifies ownership type and industry pollution as critical boundary conditions, advancing a contingency view of how cognitive drivers translate into heterogeneous ESG outcomes. Third, it unveils corporate integrity culture as a critical mediating channel grounded in social norm theory, bridging the gap between executive cognition and organizational ESG authenticity. Collectively, these insights move beyond unidimensional ESG research and illuminate the cognitive microfoundations and normative pathways essential for fostering substantive corporate sustainability.

Although knowledge spillovers fuel innovation, they are subject to significant distance-decay constraints. As the distance between regions increases, the costs of searching for, accessing, and interpreting external knowledge also rise, making effective cross-regional learning more difficult. While information and communication technologies (ICT) have developed rapidly, the specific impact of enterprise-owned social media on cross-regional knowledge flows is underexplored. By standardizing information release, these platforms may fundamentally reshape the geography of knowledge spillovers. Previous studies have examined the role of ICT in facilitating knowledge diffusion from a macro perspective, but the micro-level mechanisms through which digital disclosure platforms reshape spatial knowledge flows remain unclear. This study, therefore, contextualizes the analysis using WeChat official accounts as a representative enterprise-owned social media platform and investigates their effects on knowledge spillovers.Specifically, this study addresses three core questions: (1) whether establishing a WeChat official account significantly amplifies inter-firm knowledge spillovers; (2) whether such effects are moderated by regional absorptive capacity and multidimensional distance; and (3) how the platform′s one-to-many, weakly interactive architecture shapes exploitative versus exploratory innovation.
To answer these questions, the study uses patent citation data and constructs a multi-period difference-in-differences model to estimate the impact of firms launching WeChat official accounts on knowledge spillovers. Patent citations serve as a useful proxy for the flow and diffusion of technological knowledge, allowing researchers to trace the spatial pattern of inter-regional knowledge linkages. The analysis further investigates how the effects of enterprise-owned social media vary with geographical distance, cultural barriers, and administrative segmentation, whether these effects are influenced by regional absorptive capacity and local innovation environments, and whether enterprise-owned social media exerts heterogeneous effects on different types of technological innovation.
The results show that a firm′s patents are significantly more cited after establishing a WeChat official account, indicating a clear knowledge spillover effect. This result holds after placebo tests, robustness checks, and instrumental-variable estimations. However, the effect is spatially uneven. While WeChat official accounts can help reduce the obstacles that geographical distance and cultural barriers pose to knowledge diffusion across regions, their ability to alleviate administrative segmentation is limited. This finding suggests that while ICT can relax the time and space constraints in the circulation of information, institutional barriers continue to impose substantial frictions on inter-regional knowledge exchange. The magnitude of knowledge spillovers strongly depends on regional absorptive capacity and the local innovation environment. Cities with stronger related knowledge bases, more diversified industrial structures, or more advantageous positions in the urban hierarchy benefit more from digital knowledge spillovers. From a technological diffusion perspective, enterprise-owned social media platforms are more likely to stimulate exploitative innovation than exploratory innovation. This is because WeChat official accounts operate primarily through one-to-many communication and lack frequent reciprocal interaction; they are more suited to supporting the search, screening, and incremental application of codified or semi-codified knowledge rather than fostering exploratory technological development.
This study contributes to the existing literature in the following aspects. First, by linking firm-level digital disclosure behavior to regional knowledge flows, it provides new insights into how digital transformation reshapes the spatial patterns of knowledge diffusion. Second, by examining regional absorptive capacities and variations in the innovation environment, it clarifies the spatial heterogeneity and attenuation mechanisms of digitally mediated knowledge spillovers. Third, by highlighting the differential effects of non-interactive knowledge spillovers on exploitative and exploratory innovation, it extends innovation diffusion theory within digital media environments. Overall, the findings contribute to a deeper understanding of how enterprise-owned social media influences inter-regional knowledge transmission. The study also offers important implications for companies and policymakers aiming to foster more balanced regional innovation.

In the contemporary competitive landscape characterized by high uncertainty, innovation consortia have emerged as strategic vehicles for transcending organizational boundaries and achieving deep integration of heterogeneous knowledge. These consortia represent an advanced evolutionary form of industry-university collaboration led by enterprises and supported by universities and research institutes to tackle critical core technologies under national strategic imperatives. Unlike traditional bilateral partnerships, these consortia integrate diverse actors including horizontal competitors into complex networks, amplifying coordination costs and governance complexity. While knowledge heterogeneity holds significant innovative potential, it also triggers the "bi-decoupling" dilemma in such multi-actor environments. Bi-decoupling reflects goal-level cognitive misalignment (divergent values creating symbolic absorption) and operational-level behavioral defocus (knowledge-power gaps deflecting collaboration), constituting a governance paradox that constrains knowledge transformation. Existing research has yet to systematically elucidate how knowledge heterogeneity influences cooperative performance within the unique organizational context of innovation consortia, nor has it proposed a governance pathway for bi-decoupling from a synergistic perspective.
To examine the impact of knowledge heterogeneity and its governance pathways, this study constructs a theoretical model involving knowledge heterogeneity, knowledge integration, and cooperative innovation performance, introducing social trust level and intellectual property protection as moderating variables. Methodologically, this research employed a questionnaire survey targeting mid-to-senior managers of enterprises participating in innovation consortia listed in pilot rosters issued by science and technology authorities in Shandong, Jiangsu, Guangdong provinces, and the Yangtze River Delta region. Focusing on strategic emerging industries characterized by high technological breakthrough difficulty and intensive knowledge resource requirements, specifically biotechnology and new energy sectors, the study rigorously verified consortium membership through screening items confirming enterprise participation in leading-firm-driven consortia targeting key technologies and possessing explicit cooperative agreements. To mitigate common method bias, a two-phase longitudinal data collection approach was implemented with a three-month interval between waves. From April to July 2024, 542 questionnaires were distributed through offline channels and Wenjuanxing (a professional survey platform), yielding 411 valid responses after eliminating incomplete or invariant responses, representing a valid response rate of 75.8%. Multiple regression analysis was employed to test the hypotheses empirically.
This research integrates the Knowledge-based View and He Xie Management Theory to establish a comprehensive analytical framework. While the knowledge-based view elucidates the strategic value of heterogeneous knowledge as a critical resource for innovation, HeXie Management theory provides a dual-pathway governance mechanism through its "He Principles" framework. The Synergistic Field, activated through robust social trust levels, enables goal alignment by fostering shared value identification and psychological security among consortium members, thereby reducing cognitive fragmentation at the strategic level. Simultaneously, the Facilitative Field, implemented through rigorous intellectual property protection, establishes clear behavioral expectations and legal safeguards that suppress opportunistic behaviors and coordinate complex knowledge interactions, thus addressing operational dissonance.
The empirical results reveal a significant inverted U-shaped relationship between knowledge heterogeneity in innovation consortia and firms' cooperative innovation performance, demonstrating that an optimal level of heterogeneity maximizes cooperative outcomes through enhanced knowledge integration processes. Furthermore, social trust level significantly strengthens the relationship between knowledge heterogeneity and performance by mitigating goal-level decoupling through reinforced value identification among consortium members. Simultaneously, intellectual property protection positively moderates this relationship by reducing operational-level decoupling through effective suppression of opportunistic behaviors and providing stable institutional expectations. Both governance mechanisms work synergistically to facilitate the transformation of knowledge heterogeneity into sustainable cooperative performance, offering a comprehensive solution to the bi-decoupling challenge.