Unlock the Editor’s Digest for free
Roula Khalaf, Editor of the FT, selects her favourite stories in this weekly newsletter.
Nvidia is close to finalising a $30bn investment into OpenAI that will replace the long-term $100bn commitment agreed by the companies last year, as part of a massive new funding round for the AI start-up.
The world’s most valuable company is in the final stages of negotiations with OpenAI, and its investment could be concluded as early as this weekend, according to people with knowledge of the matter.
The $30bn equity investment forms part of a larger funding round that is on track to raise more than $100bn and will value the ChatGPT maker at $730bn, not including the new money, the people said.
OpenAI will reinvest much of its new capital into Nvidia hardware, but the companies would not proceed with the $100bn multiyear investment partnership they announced in September, the people added.
The retreat from the agreement announced to much fanfare in September comes amid investor jitters about the health of the AI sector that have helped drive US tech stocks down 17 per cent since the start of the year.
Last year’s deal, announced as a “letter of intent”, closely tied together the two companies at the heart of the AI boom — and helped to propel Nvidia above $5tn in market value a few weeks later.
It accelerated a frenzied period of dealmaking for Sam Altman’s AI start-up, which forged complex deals with rival chipmakers AMD and Broadcom and cloud providers including Oracle.
Although welcomed by stock market investors at the time, the spate of agreements tying together suppliers, customers and investors in the AI sector prompted concern among some analysts about their circular structure and a growing bubble in the space.
Under the terms of the $100bn agreement, Nvidia would have invested ten increments of $10bn as OpenAI’s demand for computing power grew over several years, in return for a significant stake in the AI start-up.
OpenAI in turn planned to buy millions of Nvidia’s AI processors as part of plans to deploy up to 10 gigawatts of new computing capacity.
But the deal never progressed from a memorandum of understanding to a formal agreement. In January, The Wall Street Journal reported the deal was “on ice”. It has now been replaced by a more straightforward arrangement in which Nvidia will invest up to $30bn in return for OpenAI stock.
That funding will support the build-out of gigawatts of new computing capacity and will probably be followed by further deals over time, according to people close to the companies.
OpenAI and Nvidia declined to comment.


Altman and Nvidia chief Jensen Huang have tried to dispel reports of cooling relations between their companies. “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,” Altman said on X earlier this month.
The Nvidia boss the next day told CNBC that any suggestion of “controversy” was “nonsense”. “We love working with OpenAI,” he said.
The San Francisco-based start-up is also in the final stages of negotiations with SoftBank, which will also invest $30bn, and Amazon, which could invest up to $50bn as part of a broader partnership involving use of OpenAI models, according to people with knowledge of the deal.
MGX, Abu Dhabi’s state-backed tech investment fund, and Microsoft are also expected to invest billions of dollars, while OpenAI executives are meeting with venture capitalists and other investors this week to drum up further interest, they added.
In those meetings, OpenAI has told investors that they intend to spend about $600bn on computing resources, including from Nvidia, Amazon and Microsoft, between now and 2030, according to one of the people.
The company has argued that access to computing resources will be the best defence against rivals, and it has made efforts to lock in as much infrastructure and supply of electrical power as possible to meet what executives anticipate will be limitless demand for AI tools.
OpenAI revenues topped $20bn earlier this year on an annualised basis, a measure favoured by start-ups that projects recent revenue over the coming year. Its revenue growth has closely tracked increases in the company’s access to computing power, with each roughly trebling each year.


