OpenAI's Sam Altman is dreaming of running 100 million GPUs in the future - 100x more than it plans to run by December 2025
TechRadarSaturday, July 26, 2025 at 12:03:00 PM
OpenAI CEO Sam Altman isn’t just thinking big—he’s thinking colossal. Right now, the company is on track to deploy over a million GPUs by the end of 2025, but Altman’s already floating an even wilder vision: scaling that up to 100 million GPUs someday. That’s a hundredfold jump, and while it’s more of a moonshot than a concrete plan, it shows just how much compute power OpenAI believes it’ll need to push AI forward. Of course, pulling that off would mean tackling huge infrastructure hurdles and eye-watering costs—so for now, it’s more of a "what if" than a "when."
Editor’s Note: This isn’t just about raw numbers—it’s a signal of how much OpenAI thinks AI’s future depends on massive computing firepower. Altman’s musings hint at a world where AI models are exponentially bigger and more capable, but they also raise practical questions: Where will all these chips come from? Who’s footing the bill? And can the energy grid even handle it? For now, it’s a bold thought experiment, but one that frames the scale of ambition (and challenges) ahead for AI’s next era.
— Curated via WP Now’s
