算力基础设施如何影响企业数字技术创新？

——基于国家超算中心建设的准自然实验

谢璇¹　崔培琳¹　张根杰¹　王运陈²

（1. 四川农业大学；2. 西华大学）

摘要：在实数融合的背景下，如何破解算力资源约束与应用门槛的双重困境，是激发企业数字技术创新活力的关键命题。本文以国家超算中心建设为准自然实验，基于2007—2023年沪深A股上市公司数据，使用多期双重差分模型，考察算力基础设施对企业数字技术创新的影响及其作用机制。研究发现，算力基础设施提升了企业数字技术创新的数量与质量，并通过政策资源供给、创新生态结构优化、数字人才集聚三条路径发挥作用。进一步研究表明，高管团队数字化战略能力会影响算力基础设施的赋能效果，且算力赋能效应在非技术密集型行业中更为明显，对大企业数字技术创新数量和小企业数字技术创新质量的促进作用更突出。本文为“算力驱动数字技术创新”提供了微观证据，也为优化算力资源布局和创新政策制定提供了政策启示。

关键词：算力基础设施；企业数字技术创新；国家超算中心；政策资源供给；创新生态结构优化；数字人才集聚

作者简介：谢璇，四川农业大学管理学院讲师，成都，611130；崔培琳，四川农业大学管理学院硕士研究生；张根杰，四川农业大学管理学院博士研究生；王运陈，西华大学管理学院教授，成都，610039。

引用格式：谢璇，崔培琳，张根杰，等. 算力基础设施如何影响企业数字技术创新？：基于国家超算中心建设的准自然实验[J]. 经济与管理研究，2026，47（7）：32-46.

How Does Computing Power Infrastructure Affect Enterprise Digital Technological Innovation?

—A Quasi-Natural Experiment Based on Construction of National Supercomputing Centers

XIE Xuan¹, CUI Peilin¹, ZHANG Genjie¹, WANG Yunchen²

(1. Sichuan Agricultural University, Chengdu 611130;

2. Xihua University, Chengdu 610039)

Abstract: Enterprise digital technological innovation is a core driver of high-quality development amid digital-real economy integration, yet firms face dual constraints of limited computing power and high application barriers. Unlike general technological innovation, digital technological innovation relies more on complex computing, large-scale data processing, and repeated experimentation. Most existing research examines the antecedents of enterprise digital technological innovation within the technology-organization-environment (TOE) framework, with insufficient attention to the unique role of computing power as a key strategic resource. This paper addresses this gap by examining whether and how computing power infrastructure empowers enterprise digital technological innovation.

Using China’s staggered National Supercomputing Center policy as a quasi-natural experiment, This paper constructs a multi-period difference-in-differences model based on data from A-share listed firms in China from 2007 to 2023. The findings show that computing power infrastructure enhances the quantity and quality of innovation, especially the latter. Mechanism tests reveal three channels: stronger policy resource supply, better innovation ecosystem structure, and greater digital talent agglomeration. Specifically, it increases local science and technology expenditure, computing-related innovation subsidies, regional digital economy development, university-industry-research collaboration, digital talent inflows, and firms’ share of technical employees. Further analysis shows that top management teams’ digital strategic capability affects this enabling role: executive digital literacy enhances both quantity and quality, whereas digital leadership mainly improves quality. Heterogeneity analysis shows stronger effects in non-technology-intensive industries, suggesting that computing power infrastructure can compensate for a lack of internal technical capabilities. Moreover, large firms gain more in innovation quantity, while small firms gain more in innovation quality. Therefore, computing power infrastructure is not only a technical public platform but also an important vehicle for reshaping the allocation of regional innovation resources, collaboration networks, and talent mobility.

This paper offers firm-level causal evidence for computing-driven digital technological innovation and shows how computing power works through several paths within a regional innovation ecosystem. It also clarifies when the effect becomes stronger or weaker. The findings offer practical recommendations, including treating computing infrastructure as a regional innovation anchor; linking policy, industry, research, and talent support; improving access to public computing services; enhancing executives’ digital strategic capability; and adopting differentiated empowerment strategies.

Keywords: computing power infrastructure; enterprise digital technological innovation; National Supercomputing Center; policy resource supply; innovation ecosystem structure optimization; digital talent agglomeration

下载全文