Here’s something you don’t hear every day: the CEO of America’s most valuable chip company is calling for more cooperation with China on artificial intelligence. Not less. More.
Jensen Huang, who runs Nvidia, recently pointed to Anthropic’s Mythos breakthrough as evidence that the US and China need to start talking about AI development together. This comes at a time when Washington is doing everything possible to keep advanced chips out of Chinese hands, and when the prevailing wisdom says we’re in a new Cold War over technology.
So what’s going on here?
The Mythos Moment
Anthropic’s Mythos represents a significant step forward in AI capabilities. For Huang, this advancement isn’t just a win for one company or one country—it’s a signal that AI is moving faster than anyone expected. And when technology moves this fast, he argues, isolation becomes dangerous.
Think about it this way: if you’re racing toward something powerful and potentially risky, do you want to be doing it blind to what the other major player is doing? Huang seems to think the answer is no.
China’s AI Sector Isn’t Standing Still
Huang has been vocal about China’s rapid progress in artificial intelligence. Despite US export restrictions on advanced chips, Chinese companies and researchers continue to make meaningful advances. The country has multiple technical parameters where it’s keeping pace or even leading, according to Huang’s assessments.
This creates an awkward reality for American policymakers. You can restrict chip exports all you want, but you can’t stop smart people from being smart. Chinese AI researchers are finding ways to do more with less, optimizing algorithms, and building their own infrastructure.
The question becomes: is it better to have some visibility into what they’re doing, or to pretend that walls will somehow slow down mathematics and computer science?
What Dialogue Actually Means
When Huang talks about US-China AI dialogue, he’s not suggesting we hand over trade secrets or abandon national interests. He’s pointing to potential mutual benefits from coordination.
Consider safety standards. If both countries are developing increasingly capable AI systems, having some shared understanding of safety protocols could prevent accidents or misunderstandings. Think of it like nuclear weapons treaties during the Cold War—you don’t have to be friends to recognize that some coordination serves everyone’s interests.
There’s also the economic angle. Nvidia sells chips globally, and China represents a massive market. But beyond Huang’s business interests, there’s a broader point: AI development requires enormous resources, and duplicating every effort on both sides of the Pacific seems wasteful.
The Uncomfortable Middle Ground
Huang’s position puts him in an interesting spot. He’s advocating for dialogue at a time when both governments are moving in the opposite direction. The US continues to tighten export controls. China pushes for technological self-sufficiency.
But maybe that’s exactly why his voice matters. Someone has to point out that the current trajectory—complete technological decoupling—might not be as smart as it sounds. Especially when the technology in question is developing faster than policy can keep up.
AI agents are already starting to handle complex tasks autonomously. The systems being built today will shape how we work, communicate, and make decisions tomorrow. Having the world’s two largest economies developing these systems in complete isolation from each other creates risks that go beyond trade disputes.
What Happens Next
Huang has indicated plans to visit China, continuing Nvidia’s engagement with the region despite political tensions. Whether his calls for dialogue gain traction with policymakers is another question entirely.
The reality is that AI development is happening whether governments coordinate or not. Chinese researchers will keep publishing papers. American companies will keep releasing new models. The question is whether there’s any communication happening between these parallel tracks, or whether we’re all just hoping for the best.
Huang’s argument, stripped down, is simple: Mythos and similar breakthroughs show that AI is advancing too quickly for isolation to be a viable strategy. You don’t have to agree with him to recognize he’s raising a question worth considering.
đź•’ Published: