Jensen Huang says he would stop selling to dangerous labs. 🤖

The question put to Jensen Huang was blunt. Sam Altman, Dario Amodei and Elon Musk all say AI should slow down. What do you see that they don't, and how can they be so wrong? His answer doesn't dodge safety. He calls it paramount, says we have agency over how this technology develops, then argues the fix is more engineering, not less. Speed up containment tools. Speed up monitoring systems. Build AIs that watch other AIs. Treating capability and safety as two separate conversations, he says, is the real mistake. He's been making this case all year. On Ezra Klein's podcast in September he put it sharper: "Nobody is building more compute today than the people asking to be slowed down." Five days later Nvidia shipped the Open Agent Safety Platform, built to let developers enforce limits on AI agents. Whatever you make of the argument, he isn't only making it in interviews. Then the interviewer asks the real question. If a lab was doing something too dangerous, would Nvidia stop selling to them? Huang calls it hypothetical and says of course. Worth knowing what that would cost him. Three customers accounted for 16%, 15% and 13% of Nvidia's revenue in the first half of this fiscal year. Close to half the company's revenue sits with three buyers it won't name. There's a catch in his position though. He's called the labs' doomsday warnings irresponsible and said there is a 0% chance AI ends the world by 2030. If that's your read of the risk, you probably never reach the day you decide a customer crossed the line. The promise is real. The bar for using it is the part nobody asked about. 🔎 We fact-checked 10 expert AI predictions so you know whose calls are worth listening to. It's free in channel bio. Follow @hyperbolic.ai for more content on AI and technology. Media: Bloomberg #ai #jensenhuang #nvidia #aisafety #technology