Enterprise & Security

Huang on 2030: 'There is 0% chance that's going to be the end of the world'

3 min read AI-generated

On September 24 Huang sits down to dinner with Xi Jinping at the White House, and he wants to talk global AI standards there. New laws, he says, are unnecessary.

Featured image for "Huang on 2030: 'There is 0% chance that's going to be the end of the world'"

Jensen Huang told CBS News exactly what he thinks of the extinction warnings of the past fortnight: nothing at all. “2030 is not going to be the end of the world. There is 0% chance that’s going to be the end of the world.” The interview with Jo Ling Kent was taped on Friday and aired Sunday on “CBS Sunday Morning”.

Who he is answering

CBS says he was responding to Jacob Coxon, the former Anthropic researcher who wrote earlier this month that developers “earnestly believe that it could kill us all by the end of the decade”. Huang calls that a set of “doomsday narratives” and says such predictions are “not grounded in science”. Then comes a line sharper than anything he has said on the subject so far: “Scaring people is unnecessary. It is irresponsible.”

The obvious conflict, asked out loud

Nvidia is the world’s most valuable company at $5.3 trillion, and it earns money in every month that training runs faster. CBS put that to him directly: why should Americans trust him, of all people, on safety?

Huang turned the argument around. “Our company’s success is directly connected to the safe deployment of products and services. If we don’t continue to do that, our value would be diminished.” Unsafe products would be bad business — so Nvidia watches out for its own sake.

The laws on the books are enough for him

Huang sees no need for new AI statutes. Product liability and the criminal provisions against unauthorised access to other people’s systems already cover what needs covering: “Apply that first — don’t let this doomsday narrative allow someone to relieve them of the laws that currently exist.” That puts him on Trump’s side. He had already called Amodei’s request for an antitrust exemption to coordinate a slowdown “completely unnecessary” last week. OpenAI disclosing six more incidents of concerning model behaviour on Thursday appears not to have moved him.

And then there is China

On September 24 Huang eats dinner with Xi Jinping at the White House. He wants to use it to discuss worldwide standards for AI development. He has little patience for export bans: “Every single chip company should go and serve the world, compete for the world.” Since December Nvidia has been allowed to sell H200 chips to approved customers in China; the strongest parts remain off limits. The only carve-outs he accepts are defence and national security.

Zero percent is not an estimate

Coxon says more than ten percent. Huang says zero. Neither number is a measurement — both are an attitude expressed in digits. The difference is that a probability of exactly zero can be neither proven nor refuted. It ends the conversation instead of having it.

What can be examined is Huang’s legal argument, and it is stronger than his critics admit. Liability law works today; an AI act works in two years. What product liability does not answer is the question everyone is actually stuck on: who is liable once nobody can say which piece of code did the thing.

Sources

NvidiaAI SafetyRegulation