The Nvidia Corporation is an American technology company which is headquartered in Santa Clara, California. Curtis Priem, Jensen Huang as well as Chris Malachowsky founded the company in 1993. The company serves a worldwide audience. Development areas are SoCs (system on a chips), GPUs (graphics processing units), APIs (application programming interfaces. These products are specifically for high-performance computing, data science as well as automotive and mobile applications.
Areas of concern for Nvidia
Nvidia NVDA.O said on Thursday its products have no “backdoors” that would allow remote access or control after China raised concerns over potential security risks in the firm’s H20 artificial intelligence chip. The Cyberspace Administration of China (CAC), the country’s internet regulator, said it was concerned by a U.S. proposal for advanced chips sold abroad to be equipped with tracking and positioning functions.
The CAC’s move cast uncertainty over the U.S. company’s sales prospects in China weeks after a U.S. export ban was reversed. The CAC was formed 14 years ago. Zhuang Rongwen is the current director who also concurrently serves as the Deputy Head of the CCP’s Publicity Department of the Central Committee. The regulator said it had summoned Nvidia to a meeting on Thursday to explain whether its H20 AI chip had any backdoor security risks, as it was worried that Chinese user data and privacy rights could be affected.
A backdoor for easy access
A backdoor risk refers to a hidden method of bypassing normal authentication or security controls. The White House and both houses of U.S. Congressย have proposed the ideaย of requiring U.S. chip firms to include location verification technology with their chips to prevent them from being diverted to countries where U.S. export laws ban sales. The separate bills and White House recommendation have not become a formal rule, and no technical requirements have been established.
Location verification specifically refers to a process whereby the physical location of individuals, activities or assets are validated or confirmed. This is done by means of geofencing, GPS tracking or other forms of location-based technologies. Nvidia has been a focus of U.S.-China relations, and China’s move comes shortly after the U.S. this monthย reversed an April banย on Nvidia selling the H20 chip to China. The company developed the H20 chip for the Chinese market after the U.S. imposed export restrictions on advanced AI chips in late 2023.
This month, Nvidia CEO Jensen Huang embarked on a public and effusive visit to China, where he sought to demonstrate his commitment to the Chinese market, met with government officials, and praised the country’s AI advances. The CAC statement did not elaborate on what backdoor security risks there could be or say what the Chinese government was considering doing as a result. Nvidia’s products are highly sought after not just by Chinese tech companies but also by Chinese military bodies, state-run AI research institutes, and universities.
China’s thoughts on the matter
The company last week placed an order with contract manufacturer TSMC for 300,000 H20 chipsets due to strong demand. In early 2023, China barred key operators of the country’s infrastructure from purchasing from Micronย MU.O, saying that a review it conducted had found the U.S. memory chipmaker’s products posed serious security risks.
Nvidia is also facing an antitrust investigation in China. The State Administration for Market Regulation announced late last year it was investigatingย the chipmakerย over suspected violations of the country’s anti-monopoly law.
The regulator said Nvidia was also suspected of violating commitments it made during its acquisition of Israeli chip designer Mellanox Technologies, under terms outlined in the regulator’s 2020 conditional approval of that deal. Nvidia’s H20 AI chip is a specialized GPU. Development was done specifically for the Chinese market, and also to comply with the export controls of the U.S. It was introduced after stricter restrictions during 2023. It does not have the same level of power as the H100. It has reduced computational density as well as interconnect speeds. Its suitability is primarily for AI inference tasks.
GCN.com/Reuters.