There's a lot of chatter these days about big AI labs designing their own chips to break free from Nvidia's grip. But Nvidia CEO Jensen Huang isn't losing any sleep over it.
During Nvidia's latest earnings call, Huang was asked about OpenAI, Anthropic, and other frontier AI labs that are reportedly working on in-house silicon. His response? He's "100% confident" they'll keep buying from Nvidia "for a very long time."
It's Not Just Chips, It's a Platform
Huang's argument is that custom AI processors are built for narrow tasks, while Nvidia offers something much bigger. "We're building something very different," he said. "These XPUs are inference specific chips for one cloud or one service. NVIDIA Corp is a platform, an entire AI factory platform that spans the entire AI life cycle."
He also pointed out that as AI services expand globally, these companies will need Nvidia's infrastructure beyond their own data centers. "I fully expect… they're going to run on NVIDIA Corp all around the world," Huang said, adding that he has "100% confidence" the frontier AI labs "are going to be customers and partners of ours for a very long time."
Huang also framed Nvidia's investments in these AI companies as a smart long-term bet, not a risk. "Investing in these companies are once in a generation opportunity," he said. "The only regret that I have is that I didn't invest more and sooner."
This all comes as OpenAI, Anthropic, and others push forward with custom chip programs to reduce their reliance on third-party hardware for certain workloads. But Huang's message to investors is clear: proprietary inference chips and Nvidia's end-to-end AI platform can coexist, and the world's biggest AI labs are likely to keep buying both.