Supercomputing “Engine”

Taking the operation of supercomputing centers as a quasi-natural experiment, this paper adopts the double machine learning model to systematically examine the impact and mechanism of computing power infrastructure on urban key digital technology innovation. This study finds that computing power infrastructure significantly enhances the key digital technology innovation. Mechanistically, it facilitates such innovation indirectly through two major channels. It stimulates innovation vitality, accelerates digital talent flow and promotes digital technology application via sound market mechanisms. Meanwhile, it facilitates policy support for digital innovation, strengthens digital intellectual property protection and facilitates data element agglomeration through capable government mechanisms. The promotional effect of computing power infrastructure on key digital technology innovation is more prominent in cities featured with inferior business environment, inadequate sci-tech talent agglomeration and sluggish market potential, as well as those with low public data openness, excessive administrative redundancy and high fiscal decentralization. The influence of computing power infrastructure on key digital technology innovation shows a distinct long-distance leapfrog effect from “origins” to “transformation sites”, which can effectively bridge the gap in key digital technological innovation across regions. This study provides valuable implications for exerting the positive functions of effective markets and capable governments in digital economy development, and facilitates the in-depth integration of computing power infrastructure construction and key digital technology innovation.