Semiconductor giant Nvidia reasserted the security of its AI processors in a blog post on Tuesday, pointing out that inserting weaknesses into chips would “undermine global digital infrastructure and fracture trust in US technology”.
“There are no back doors in Nvidia chips. No kill switches. No spyware,” the post said. “That’s not how trustworthy systems are built – and never will be.”
According to the post written by Nvidia chief security officer David Reber Jnr, graphics processing units (GPUs) from the company, which has been designing chips for more than 30 years, are “used across industries – from healthcare and finance to scientific research, autonomous systems and AI infrastructure”.
“Embedding back doors and kill switches into chips would be a gift to hackers and hostile actors,” the post said. “Established law wisely requires companies to fix vulnerabilities – not create them.”
Nvidia’s blog post reiterated the company’s strong denial of alleged security risks in its H20 GPUs, which prompted the Cyberspace Administration of China (CAC) last week to conduct an inquiry into the company.
“Nvidia does not have ‘back doors’ in our chips that would give anyone a remote way to access or control them,” a company representative said in an email to the South China Morning Post, a day after the CAC summoned the US firm to discuss its security concerns.
