Nvidia has announced plans to invest up to $100 billion (€85 billion) in OpenAI in return for a significant stake in the ChatGPT maker, as part of a joint effort to build data centres for artificial intelligence.
OpenAI plans to buy millions of Nvidia’s AI processors as part of a deal to deploy up to 10 gigawatts of capacity, the two companies said on Monday, the equivalent output of 10 nuclear reactors.
Should Nvidia commit the full amount, it would be the largest-ever investment in a private company, reflecting the vast resources the $4.5 trillion chip giant wields as the world’s most valuable company.
The contract ensures that Nvidia stays at the heart of OpenAI’s products for years to come.
RM Block
The deal could generate hundreds of billions of dollars in revenue for Nvidia, whose shares were up about 4 per cent at market close on Monday. But the groups did not specify a timeframe for the deployment and many of the details remain uncertain.
“Everything starts with compute,” said Sam Altman, chief executive of OpenAI. “Compute infrastructure will be the basis for the economy of the future, and we will utilise what we’re building with Nvidia to both create new AI breakthroughs and empower people and businesses with them at scale.”
Jensen Huang, Nvidia chief executive, called the investment and infrastructure partnership “the next leap forward” in AI.
Nvidia plans to buy equity in OpenAI progressively over time as its systems are deployed. The investment would be made in cash, said one person familiar with the deal.
The first $10 billion will be invested when the first gigawatt is deployed. At OpenAI’s current $500 billion valuation, that would give Nvidia about 2 per cent of the AI start-up. Later investments would be made at the prevailing valuation.
Over the course of the build-out, OpenAI planned to spend more than $100 billion on Nvidia chips, one person familiar with the deal said. That side of the deal could in part be funded by Nvidia’s investment, with future revenue growth and other investments to cover the remainder.
The agreement is the latest in a string of megadeals struck by lossmaking OpenAI in recent months, including an agreement to purchase $300 billion in computing power from Oracle over five years and a plan to start producing its own AI chips with Broadcom.
OpenAI is also in complex negotiations with its earliest backer, Microsoft, over converting to a new corporate structure that would allow investors to own equity in the business for the first time.
All of the 10GW in infrastructure planned in the deal relates to previously unannounced projects, and would involve 4mn to 5mn of Nvidia’s AI chips. Most of it is expected to be deployed in the US, one source close to the deal said.
Nvidia estimates that the total cost of 10GW of AI infrastructure would require an investment of as much as $400 billion, incorporating the chips, land and physical infrastructure.
The infrastructure deal, the first phase of which is planned to come online in the second half of 2026, will deploy Nvidia’s next-generation “Vera Rubin” chip system, the successor to its current generation Blackwell technology.
If completed in full, Nvidia’s investment would be roughly equivalent to analysts’ projections of the company’s free cash flow for the current financial year.
“It’s an entirely feasible investment here from Nvidia: they have a very strong cash flow, a very good balance sheet, and this is a company that quite frankly is just churning out growth,” said Melissa Otto, head of research at S&P Global Visible Alpha.
A $100 billion investment would surpass the upfront value of any of the tech industry’s largest M&A transactions to date, including Microsoft’s $75 billion purchase of Activision Blizzard and Broadcom’s $69 billion VMware deal.
However, it was hard to compare the precise value of the deal to others in the sector, Otto said, because of the uncertain timing and cadence of the investment.
Dimitri Zabelin, AI analyst at PitchBook, said the deal was the latest example of how “Nvidia is consolidating control over the AI stack and reinforcing its position as the [sector’s] indispensable enabler”.
Nvidia has become a prolific investor in AI companies that rely on its chips, including upstart cloud computing providers such as CoreWeave and Nebius, application developers xAI and Perplexity, and robotics start-ups Figure and Wayve.
But Michael Cusumano, professor of technological innovation and entrepreneurship at MIT’s Sloan School of Management, said Nvidia’s investment was “kind of a wash” since “Nvidia is investing $100 billion in OpenAI stock and OpenAI is saying they are going to buy $100 billion or more of Nvidia chips”. - Copyright The Financial Times Limited 2025
