We don’t need AI regulation — leave safety to us, Nvidia’s Jensen Huang says

At Salesforce’s Dreamforce conference on Tuesday, Nvidia founder and CEO Jensen Huang made his case plainly: AI is not an alien mind, just human-built hardware and software. In his view, that makes safety an engineering problem, not a legal one. Existing laws should be enough, he argued, and new AI-specific regulation is unnecessary.

Huang pushed the idea that market pressure will do most of the work. If a product or service isn’t safe or reliable, companies should simply hold back and keep testing until they’re confident it’s ready. Speed and safety, he said, are not enemies. They can coexist if companies know when to pause.

That’s a neat theory, and it’s also exactly the sort of theory that sounds better when your company is riding the AI boom. Nvidia has been building the hardware behind AI since before ChatGPT existed, and Huang now describes a world where the company’s ambition rises with the technology. He was blunt about that too: “The sky’s the limit.” For Nvidia, for every industry, for every country.

But software history is not kind to the “trust us” school of safety. The source points to the 2024 CrowdStrike outage, which knocked flights off course and caused business havoc, as a reminder that even ordinary software can misbehave at scale. It also cites Meta’s $18 billion settlement over child-safety harms, along with AI-related lawsuits and the Hugging Face hacking incident. In other words, bad outcomes do not require bad intentions.

Huang’s bet is that existing product liability law, plus company judgment, is enough. Maybe. But the moment a system can cause harm across borders and industries, waiting for courts to catch up starts to sound less like prudence and more like a gamble with everyone else’s chips.