OpenAI and Nvidia pursue $250B guarantee for 10 GW Ohio AI campus

If this guarantee closes anywhere near the reported terms, it would be one of the clearest signals yet that the race to AGI is now constrained more by capital and energy than by algorithms. A 10 gigawatt campus is an order of magnitude beyond today’s largest AI clusters. It implies planning for trillions of model parameters, persistent agentic workloads and always‑on simulation environments, not just occasional inference. For OpenAI, owning or controlling this much dedicated compute would be a decisive break from dependence on Microsoft and other cloud landlords, giving it a long runway to train successive frontier models on its own schedule.

For Nvidia, a 250 billion dollar contingent guarantee tied to a single customer would crystallize the circular financing dynamic already visible in AI: chip vendors funding the very demand that drives their revenue. That deepens systemic risk but also entrenches Nvidia as the default hardware substrate for frontier labs. Strategically, this campus would anchor the United States in the global compute race just as Chinese players like Moonshot, DeepSeek and CXMT are ramping their own infrastructure. Whoever controls multi‑gigawatt‑scale campuses like Piketon will set the pace on training the first systems that plausibly approach AGI.