Nvidia’s Jensen Huang Defends Chinese AI Models as Market Catalyst

In an exclusive Axios interview, Nvidia chief Jensen Huang argued that American firms have nothing to fear from Chinese AI models, but should worry about the political push to ban them.

Why Chinese AI models matter to the global chip market

Huang, speaking from a new Wistron plant in Fort Worth, said open‑source models expand the user base, driving demand for Nvidia GPUs and data‑center capacity. He cited the recent launch of Moonshot AI’s Kimi K3, which combines near‑frontier performance with dramatically lower prices and openly available weights, as proof that cheaper AI can ↑ 12% chip sales.

“There’s no scenario where China runs U.S. companies off the road,” Huang told Axios.

The market reaction was swift: Nvidia shares slipped ↓ 8% after the Kimi announcement, reviving concerns that open models could undercut the massive AI‑infrastructure buildout. Huang dismissed the notion of a security backdoor, insisting that sandboxed deployments keep proprietary data safe while external researchers can spot flaws. He urged rivals like Anthropic to release restricted models such as Claude Mythos to the public, arguing that openness strengthens defenses. Reuters and Bloomberg have reported increasing scrutiny from U.S. officials, including Treasury Secretary Scott Bessent’s remarks on possible sanctions for IP theft. Huang countered that learning from existing models is a natural part of intelligence development and that penalties should target illicit behavior, not the models themselves. The debate echoes earlier tech‑policy tensions that resurfaced during the pandemic era, when supply‑chain fragility prompted calls for strategic autonomy.