At almost the same moment Nvidia CEO Jensen Huang was reacting to OpenAI’s surprising partnership with AMD, Sam Altman was already teasing more major deals ahead.
Huang, appearing on CNBC’s Squawk Box, admitted he had little knowledge of the AMD announcement, which gives OpenAI the chance to acquire up to 10 percent of AMD’s stock in exchange for helping develop its next-generation AI chips. The deal effectively makes OpenAI a shareholder in the chipmaker.
Nvidia’s arrangement with OpenAI takes the opposite form. After agreeing to invest as much as $100 billion into the AI giant, Nvidia has become one of OpenAI’s biggest backers, supplying it directly with advanced hardware for the first time.
Huang said the new sales approach will allow OpenAI to build its own data centers, describing it as a step toward becoming a “self-hosted hyperscaler.” But he admitted the cost is massive—each gigawatt of AI infrastructure could run between $50 billion and $60 billion.
In 2025 alone, OpenAI has commissioned 10 gigawatts of U.S. facilities through its $500 billion Stargate project with Oracle and SoftBank, alongside a $300 billion cloud deal with Oracle. Its partnerships with Nvidia and AMD add another 16 gigawatts of capacity, while expansion continues in the U.K. and across Europe.
During an interview on Andreessen Horowitz’s a16z Podcast, Altman said these deals are only the beginning. Confident in OpenAI’s upcoming models and technology roadmap, he hinted that the company will pursue even more collaborations in the coming months.
“To make the bet at this scale, we need the whole industry to support it,” Altman said, noting that OpenAI’s infrastructure race will rely on deep partnerships “from electrons to model distribution.”
OpenAI’s billion-dollar momentum, it seems, is only just getting started.
