Nvidia's $100 Billion OpenAI Deal Stalls as Huang's Doubts Grow

A $100 billion partnership between Nvidia and OpenAI, once heralded as the largest computing project ever conceived, has effectively stalled. The non-binding memorandum of understanding, announced last September with Nvidia CEO Jensen Huang promising to build vast computing power for the AI lab, never progressed to a final contract. Sources familiar with the matter say the deal has quietly unraveled, a significant retreat from the transformative alliance both companies publicly celebrated.

Behind the scenes, Huang has expressed growing skepticism about the arrangement to industry associates. He has pointed to what he sees as a lack of business discipline at OpenAI and voiced concerns about the company's intensifying competition with Google's Gemini and Anthropic's Claude. These private reservations contrast sharply with the public optimism that sent Nvidia's stock soaring last fall.

Rather than a complete break, discussions have shifted to a far more modest alternative: a potential equity investment by Nvidia worth tens of billions of dollars into OpenAI's current funding round. This scaled-back approach reflects a recalibration by the chip giant, which remains strategically motivated to support a top customer but is unwilling to shoulder the extraordinary risk of the original proposal, which included discussions of Nvidia guaranteeing loans for OpenAI data centers.

For OpenAI CEO Sam Altman, the collapse is a setback. His strategy of announcing massive computing deals to secure capacity for next-generation AI models is now under scrutiny. The company, laying groundwork for a potential 2026 public offering, must demonstrate it can secure reliable infrastructure without the guaranteed backing of the Nvidia megadeal. The episode reveals the precarious balancing act in AI, where astronomical ambitions meet the hard realities of binding contracts and business fundamentals.