Nvidia has found itself at the center of escalating technological tensions between the United States and China. Chinese authorities formally summoned Nvidia executives to address concerns about potential “backdoor” security vulnerabilities in the company’s H20 artificial intelligence chips, specifically developed for the Chinese market. The Cyberspace Administration of China (CAC) announced it had interviewed Nvidia officials, demanding explanations and supporting evidence related to these suspected vulnerabilities.
Regulators in Beijing allege that Nvidia’s H20 AI chips, which form part of a China-specific product line, may contain features that enable tracking, pinpointing, or remote shutdown of the chips. Such capabilities, they argue, could be used for foreign monitoring or even the disablement of critical Chinese systems—raising profound concerns over data sovereignty and national security.
The H20 chips recently re-entered the Chinese market after the U.S. government relaxed export restrictions initially imposed in prior years. Nvidia developed these chips to comply with U.S. export controls by reducing certain performance metrics while retaining appeal for Chinese customers. This latest regulatory scrutiny in China comes amid proposals in the U.S.—notably provisions in the proposed U.S. Chip Security Act—that would require export-bound advanced semiconductors to have tracking or remote shutdown features, fuelling additional mistrust in China.
In response, Nvidia firmly denied the presence of any intentional “backdoors” or features that would allow outside entities to access, track, or control its chips without authorization. A company spokesperson stated, “Cybersecurity is of utmost importance to us. Nvidia does not incorporate ‘backdoors’ in our chips that would enable anyone to access or control them remotely.”