OpenAI and Nvidia have signed a letter of intent for a $100 billion partnership that could reshape how AI systems are trained and run. Under the plan, at least 10 gigawatts of Nvidia hardware would back OpenAI’s next-generation infrastructure, intended to train and operate future models that the company says are aimed at superintelligence.
Nvidia said it intends to invest up to $100 billion in OpenAI as those systems are brought online. The first phase of the rollout is slated for the second half of 2026 and will be powered by Nvidia’s upcoming Vera Rubin platform.
The arrangement highlights how tightly linked the largest players in the industry have become. Nvidia, the dominant supplier of AI accelerators, would pick up a financial stake in OpenAI, one of its biggest customers. OpenAI would receive a large capital infusion plus secured access to vast quantities of Nvidia processors that are in high demand across the sector.
Some rivals and market watchers reacted uneasily. The deal could strengthen Nvidia’s position in chips while reinforcing OpenAI’s lead on models and software, feeding concerns about competition and market concentration inside a small group of firms.
A person familiar with the matter described the partnership as a two-step structure: Nvidia will acquire non-voting shares in OpenAI, and OpenAI will then use the proceeds to purchase Nvidia systems. That flow of funds and hardware ties the companies’ financial interests to the procurement cycle for compute.
“Everything starts with compute,” OpenAI CEO Sam Altman said. “Compute infrastructure will be the basis for the economy of the future, and we will utilize what we’re building with Nvidia to both create new AI breakthroughs and empower people and businesses with them at scale.”
Corporate spokespeople said finer points of the partnership will be finalized in coming weeks. They pointed out that a 10-gigawatt deployment would draw roughly the same amount of electricity as more than 8 million U.S. households.
Financial markets reacted quickly. Nvidia shares surged, rising as much as 4.4% to a record high after the announcement. Oracle, which is collaborating with OpenAI, SoftBank and Microsoft on a $500 billion global AI data-center initiative called Stargate, jumped about 6%.
According to another person familiar with talks, after the final pact is signed OpenAI will formally buy Nvidia systems. Nvidia would then make an initial equity investment of about $10 billion in OpenAI, which was last valued at roughly $500 billion.
Nvidia’s first hardware shipments tied to the agreement are expected to arrive in late 2026. The companies plan to bring one gigawatt of the promised computing power online in the second half of that year using the Vera Rubin platform.
Industry analysts gave guarded praise for the deal’s ambition while flagging potential circularity in its economics. “On the one hand this helps OpenAI deliver on what are some very aspirational goals for compute infrastructure, and helps Nvidia make sure that that stuff gets built. On the other hand the ‘circular’ concerns have been raised in the past, and this will fuel them further,” said Stacy Rasgon, an analyst at Bernstein.
OpenAI has explored other routes to reduce reliance on third-party chips. The company, like Google and Amazon, has been working on custom designs and partnerships to lower costs and add flexibility. A person close to OpenAI said the Nvidia arrangement does not alter existing compute plans, including its ongoing collaboration with Microsoft.
Earlier reporting by Reuters said OpenAI had been working with Broadcom and Taiwan Semiconductor Manufacturing Co. on chip designs. After news of the Nvidia tie-up, Broadcom shares slipped about 0.8%.
OpenAI has expanded rapidly, now reporting more than 700 million weekly active users and wide adoption by businesses and developers. Backers of the Nvidia deal say the scale of the compute commitment could accelerate OpenAI’s push toward artificial general intelligence.
The pact joins a string of high-profile industry alliances. Microsoft has poured billions into OpenAI since 2019. Nvidia recently announced a chip collaboration with Intel and pledged $5 billion in support for related initiatives. Nvidia also participated in OpenAI’s $6.6 billion financing round in October 2024.
Given the magnitude of the investment and the strategic links, the agreement may attract scrutiny from competition regulators. Last year, the U.S. Justice Department and Federal Trade Commission reached an understanding that allows for closer review of the roles played by Microsoft, OpenAI and Nvidia in the AI market. So far, the Trump administration has favored a lighter regulatory touch than the Biden administration when it comes to competition matters.
This month, OpenAI and Microsoft said they had signed a non-binding agreement to convert OpenAI into a for-profit company, a shift that would bring changes to governance and capital structure.
“The deal could change the economic incentives of Nvidia and OpenAI as it could potentially lock in Nvidia’s chip monopoly with OpenAI’s software lead. It could potentially make it more difficult for Nvidia competitors like AMD in chips or OpenAI’s competitors in models to scale,” said Andre Barlow, an antitrust lawyer at Doyle, Barlow & Mazard.
Barlow added that the Trump administration’s pro-business posture on regulation has removed some barriers that might otherwise slow the growth of advanced AI projects.