Nvidia is nearing a revised investment in OpenAI, significantly scaling back its initial commitment to $30 billion, replacing a previously discussed $100 billion deal. The move comes amid growing investor anxieties surrounding the artificial intelligence sector and questions about the pace of technological advancement at OpenAI.
The revised agreement, expected to be finalized as early as this weekend, is part of a larger funding round aiming to raise over $100 billion for the ChatGPT maker, valuing OpenAI at $730 billion before the infusion of new capital, according to people familiar with the matter.
The original $100 billion agreement, announced in September 2025 as a non-binding letter of intent, envisioned Nvidia investing in ten $10 billion increments as OpenAI’s computing needs expanded. This would have tied the two companies closely together, with OpenAI purchasing millions of Nvidia’s AI processors and requiring up to 10 gigawatts of new computing capacity – equivalent to the output of roughly ten nuclear reactors.
However, the deal stalled, and Nvidia CEO Jensen Huang has since characterized the $100 billion figure as “never a commitment.” The retreat reflects a broader cooling in the initial exuberance surrounding AI investments, with US tech stocks experiencing a 17 percent decline since the start of the year.
The shift also comes as OpenAI explores alternative chip suppliers, reportedly dissatisfied with the speed of some Nvidia chips for inference tasks – the process by which AI models generate responses. According to reports, OpenAI staff attributed performance limitations in its Codex AI code-generation tool to Nvidia’s GPU-based hardware. OpenAI CEO Sam Altman publicly reaffirmed his company’s relationship with Nvidia earlier this month, stating on X, “We love working with Nvidia and they make the best AI chips in the world. We hope to be a gigantic customer for a very long time.” Huang echoed this sentiment, dismissing reports of friction as “nonsense.”
The new $30 billion investment from Nvidia will support the expansion of OpenAI’s computing capacity. OpenAI is also in discussions with other investors, including SoftBank, which is considering a $30 billion investment, and Amazon, potentially investing up to $50 billion as part of a broader partnership. MGX, Abu Dhabi’s state-backed tech fund, and Microsoft are also expected to participate in the funding round.
OpenAI has reportedly told potential investors that it anticipates needing approximately $600 billion in computing resources between now and 2030, relying on Nvidia, Amazon, and Microsoft to provide the necessary infrastructure. The company views access to computing power as a critical competitive advantage and is actively seeking to secure long-term supply agreements and sufficient electrical power capacity.
OpenAI’s revenue has grown rapidly, exceeding $20 billion on an annualized basis, mirroring the increase in its access to computing resources. The company’s revenue and computing capacity have both roughly tripled each year.
The original agreement between Nvidia and OpenAI, while initially welcomed by investors, also raised concerns about the circular nature of the dealmaking within the AI sector. The flurry of agreements between suppliers, customers, and investors prompted some analysts to question whether a bubble was forming.
The revised investment structure represents a more straightforward equity investment, with Nvidia acquiring stock in OpenAI in exchange for capital. This shift suggests a recalibration of expectations and a more cautious approach to the massive infrastructure build-out initially envisioned. While the long-term implications remain to be seen, the scaling back of the initial $100 billion commitment signals a more pragmatic phase in the development of the AI ecosystem.
