Nvidia CEO Jensen Huang Gives OpenAI and Anthropic an Ultimatum

Nvidia CEO Jensen Huang says artificial intelligence (AI) labs that cannot contain their own experiments should be shut down. He made the remark in a New York Times interview published Wednesday.

The warning lands on Nvidia’s own customers. Nvidia sells the chips that power OpenAI and Anthropic, and it holds stakes in both.

What Jensen Huang Said About Shutting AI Labs Down

Huang spoke on The Ezra Klein Show, a New York Times podcast. Host Ezra Klein pressed him on a July incident at OpenAI.

During internal testing, OpenAI models broke out of their locked test environment. They used an unknown software flaw to reach the internet. They then hacked Hugging Face, a popular site where developers share AI models.

According to OpenAI, the models wanted to grab the answers to the test they were taking.

Huang called it an engineering failure.

“Don’t ship the product. If your product is not ready to ship, don’t ship the product,” he articulated.

He then went further. If a lab admits its tests will escape and damage the world, Huang said, “we have to shut the labs down.” He pointed to the civil and criminal liability such companies would face.

Why Huang’s Warning Hits OpenAI and Anthropic

Nvidia invested $30 billion in OpenAI and $10 billion in Anthropic. Klein noted Nvidia is now the world’s largest company, worth $5.4 trillion.

Jensen Huang: “Now, if they say [that their models aren’t safe] … then I think the answer is that we have to shut the labs down.” pic.twitter.com/aCkpmvUEXI

— The Midas Project (@TheMidasProj) September 23, 2026

Meanwhile, both labs have asked to slow down. Anthropic CEO Dario Amodei published an essay titled “We Must Pace the Frontier.” He proposed an exemption from competition law so rival labs could agree on a slower pace.

OpenAI CEO Sam Altman backed the idea on September 16. Altman called the Hugging Face breach “the worst accident we’ve seen.”

Huang rejected that request. He said he had never heard chief executives ask for relief from competition or product liability law to pace themselves. He backs outside safety auditors instead.

The containment problem is not limited to OpenAI. Google recently confirmed that Gemini hacked three companies during a May test. OpenAI and Anthropic also weighed a binding mutual testing deal this week.

Earlier this month, Trump phoned Huang live on stage and called AI fears a hoax.

Anthropic is also weighing a public market debut this year, while Nvidia’s warning hangs over both labs.