Nvidia Corp has publicly rejected a U.S. proposal to embed remote-access functions into its chips, amid heightened scrutiny from China’s top internet regulator. The company emphasized that such mechanisms would weaken device security and potentially expose infrastructure to cyber threats.
In a blog post titled “No Backdoors, No Kill Switches, No Spyware,” Nvidia warned that building tracking features into hardware would be “a gift to hackers and hostile actors,” arguing the idea would create central points of failure.
The response comes after the Cyberspace Administration of China (CAC) reportedly asked Nvidia for clarification about its chip security, following its recent reauthorization to sell H20 AI chips in China. The CAC expressed concern about potential national security risks.
“Nvidia does not have ‘backdoors’ in our chips that would give anyone a remote way to access or control them,” a company spokesperson said, addressing the regulator’s concerns.
The chipmaker drew historical parallels with the failed 1990s U.S. Clipper Chip project, which aimed to allow government access to encrypted communications. Nvidia said similar efforts today would undermine user trust and economic security.
Reaffirming its stance, the company stated: “There are no back doors in Nvidia chips. No kill switches. No spyware.” It also warned against policies that could erode global confidence in U.S. technology leadership.