Nvidia bets on billion investment in OpenAI – A high-stakes game

Nvidia plans to invest up to $100 billion in OpenAI while simultaneously supplying millions of AI chips to the creator of ChatGPT. This unprecedented strategy includes a non-voting stake in one of its largest customers. OpenAI will be provided with part—but not all—of the capital needed to develop next-generation supercomputers. Nvidia’s initial investment of $10 billion is set to flow into one gigawatt capacity, with construction using the new Vera Rubin chips set to begin in the second half of 2026. However, questions regarding financing remain open. Nvidia’s CEO Jensen Huang mentioned in an earnings call in August that construction costs are about $50 billion per gigawatt. Of this, $35 billion is expected to flow into Nvidia’s equipment, but there is uncertainty about the additional financing. OpenAI did not comment on the financing plans, and Nvidia refrained from further comments. OpenAI, originally a non-profit organization, is seeking a restructuring to a for-profit corporation. Talks with major investor Microsoft resulted in a preliminary plan to convert to a Public Benefit Corporation, which requires regulatory approval. Meanwhile, OpenAI's valuation remains unclear at $500 billion, alongside the potential impacts of Nvidia’s investment. Also uncertain are the effects on competitors like Anthropic and Microsoft, who also seek Nvidia’s coveted chips, and AMD’s possible sales opportunities. Oracle, which had announced significant cloud contracts with OpenAI, could find itself in a safe position due to Nvidia’s commitment. Despite questions about OpenAI's financial capability, Nvidia’s willingness to invest strengthens the basis for Oracle’s forecasts.