NVIDIA’s plan to invest $100 billion in OpenAI may be running into trouble. The latest reports say that the partnership—announced with great fanfare last September—has now fallen into a stalemate.

According to the memorandum of understanding previously announced by both sides, NVIDIA would provide up to $100 billion to support OpenAI in building at least 10 gigawatts of computing capacity, with OpenAI leasing NVIDIA chips as part of the arrangement. However, people familiar with the matter said the negotiations never entered a substantive phase, and there are clear internal doubts at NVIDIA about the transaction.

This development could have a significant impact on the AI industry. NVIDIA is currently the dominant force in AI compute, while OpenAI is one of the most prominent players on the model side. If the two were deeply tied together, they could have formed a strong closed-loop ecosystem. But with the plan now on hold, it highlights differences between the two companies in risk assessment, commercial timing, and the competitive landscape.

Last September, NVIDIA and OpenAI jointly announced the cooperation at NVIDIA’s headquarters in Santa Clara, calling it “the largest computing project in history.” On the day of the announcement, NVIDIA’s shares rose nearly 4%, and its market capitalization briefly approached $4.5 trillion, fully igniting market sentiment. OpenAI had expected to finalize formal negotiations within a few weeks—but months have passed, and the agreement has not even made it through the early stages.

Inside NVIDIA, some people have begun questioning the scale and risk of the commitment: $100 billion was a substantial portion of NVIDIA’s market value at the time. While the funds would ultimately flow back into purchasing NVIDIA’s own GPUs—seemingly “money moving from the left hand to the right hand”—the investment could still become a sunk cost if OpenAI’s spending spirals out of control, if it falls behind in model iteration, or if its IPO plans run into trouble.

Both sides are now re-evaluating the future direction of the partnership. Recent discussions reportedly include the possibility of NVIDIA making an equity investment of tens of billions of dollars in OpenAI’s current funding round.

If the $100 billion order truly stops, then the shift would essentially downgrade the relationship—from a strategically deep binding partnership to simply becoming one of OpenAI’s major shareholders. The core issue behind the stalled deal is not technology, but trust and expectation management. NVIDIA worries about being taken for a ride, while OpenAI needs to secure computing capacity quickly. But after a high-profile announcement failed to materialize, the market has begun questioning the reliability of the commitment.

Market interpretation:

NVIDIA itself announced in November that it would invest up to $10 billion in Anthropic, which in itself serves as a form of hedging. In private conversations with industry insiders over recent months, Jensen Huang has reportedly expressed concerns about OpenAI’s commercial discipline, while also raising worries about the risk of OpenAI falling behind in the competition.