DarkReading.com reported that “The US banned the sale of AI chips to China and then backed off. Now, Chinese sources are calling on NVIDIA to prove its AI chips have no backdoors. Chipmakers Nvidia and Advance Micro Devices continue to face challenges in shipping advanced AI processors to China, with both national security and hardware security at issue for both China and the US.” The August 13, 2025 article entitled “China Questions Security of AI Chips From NVIDIA, AMD” (https://www.darkreading.com/cyber-risk/china-questions-security-ai-chips-nvidia-amd) included these comments:
In 2022, the US government placed a ban on exporting high-end AI chips, until last month, when the Trump administration decided to allow the shipment to China of less sophisticated processors, as long as manufacturers pay the US government a 15% fee. Now, however, China has demanded that the most significant chipmaker, NVIDIA, prove that its processors do not have exploitable security flaws or backdoors, with Chinese state media claiming that NVIDIA’s H2O chips are not safe.
The Cyberspace Administration of China called for the company to address risks that its H2O chips have backdoor capabilities, according to China Daily, an English-language news outlet owned by the Central Propaganda Department of the Chinese Communist Party.
“If NVIDIA’s chips really have backdoor risks, that will become its ‘self-dug grave,'” the state media outlet quoted Pan Helin, a member of the Ministry of Industry and Information Technology’s Expert Committee for Information and Communication Economy. “Users, not just companies in China, but clients across the world, may abandon its chips over fears of remote shutdowns or data theft.”
Backdoors — and claims of backdoors — are nothing new. Both the US and China have been targeted by — and deployed — backdoors in technology to aid in espionage. Both countries have also used the fear of backdoors to support policies, such as recent US concerns over the security of systems used in US ports and claims by China that its earthquake monitoring systems had been targeted by US threat actors.
Yet private US companies, especially cybersecurity firms, have rarely cooperated with such schemes. Nvidia would never add any backdoor code to its chips, chief security officer David Reber Jr. said in an Aug. 5 statement.
“There are no backdoors in Nvidia chips. No kill switches. No spyware,” he wrote. “That’s not how trustworthy systems are built — and never will be.”
I’m very concerned, what about you?
First published at https://www.vogelitlaw.com/blog/china-is-questioning-security-of-ai-chips-from-nvidia-and-amd