On July 31, China's Cyberspace Administration (CAC) summoned U.S. chip giant NVIDIA for a regulatory meeting, citing serious cybersecurity concerns over the company's H20 AI accelerator chip sold in China. The move follows recent reports suggesting the chip may contain backdoors and vulnerabilities that pose risks to network and data security.

The CAC stated that, under China's Cybersecurity Law, Data Security Law, and Personal Information Protection Law, NVIDIA is required to provide a detailed explanation of the alleged security flaws and submit relevant supporting materials.

The development comes amid growing concerns in China over potential foreign surveillance capabilities embedded in imported high-performance chips. Several U.S. lawmakers, led by Senator Tom Cotton, introduced the "CHIP Act" in May, which would mandate U.S. AI chip exports—including those from NVIDIA—to be equipped with tracking, geolocation, and remote shutdown features. According to American AI experts, such technologies have already been developed and integrated into some advanced chips.

The legislation is part of broader U.S. efforts to assert tighter control over the global use of its semiconductor technology, aiming to prevent strategic competitors from utilizing U.S.-designed chips in sensitive applications.

In response to the CAC's inquiry, an NVIDIA spokesperson said on July 31 (local time), “Cybersecurity is critically important to us. NVIDIA chips do not contain backdoors and cannot be accessed or controlled remotely by anyone.”

The incident highlights rising geopolitical and regulatory scrutiny surrounding advanced AI hardware, as nations work to protect data sovereignty while managing escalating tech tensions.