Stay informed with free updates
Simply sign up to the Artificial intelligence myFT Digest — delivered directly to your inbox.
Mira Murati’s Thinking Machines Lab has struck a chip supply deal with Nvidia worth tens of billions of dollars, as the former OpenAI chief technology officer accelerates its push to build AI tools for businesses.
The San Francisco-based start-up on Tuesday said it would deploy at least one gigawatt of Nvidia’s next-generation Vera Rubin chips in a multiyear partnership with the US chip giant.
The deal also includes a further “significant” investment from Nvidia, which previously participated in Thinking Machine’s $2bn seed round that valued it at $10bn last year. The companies declined to provide details of the new investment.
The new Thinking Machines deal is the latest in a series of investments by Nvidia in some of the largest customers and consumers of its AI chips.
Such funding arrangements with Nvidia’s own customers have drawn concerns about circular financing in the industry, as the $4.4tn semiconductor behemoth deploys its massive cash reserves into its customer ecosystem.
“Nvidia’s technology is the foundation on which the entire field is built,” said Murati. “This partnership accelerates our capacity to build AI that people can shape and make their own, as it shapes human potential in turn.”
Founded by Murati a year ago, Thinking Machines’ flagship product, Tinker, allows enterprises to fine-tune and customise large language models for their particular business applications without needing to build and manage complex AI training infrastructures.
Nvidia chief executive Jensen Huang said he was “thrilled to partner” with Thinking Machines to help it “realise its exciting vision for the future of AI”.
The first phase of the multiyear deal is targeted for the start of next year. Huang has previously said that one gigawatt of AI data centre capacity costs around $50bn to $60bn, with Nvidia’s products accounting for around $35bn of that cost.
A multiyear, $100bn “strategic partnership” with OpenAI, announced in September, was recently abandoned in favour of a $30bn equity investment by Nvidia. Huang said last week that the deal “might be the last time” Nvidia invests in OpenAI before a potential IPO later this year.
As an executive of OpenAI, Murati worked on the development of products such as ChatGPT, the Dall-E image generator and its voice mode. She briefly replaced Sam Altman as interim chief executive officer during a board coup in November 2023.
The Nvidia deal comes following upheaval within Thinking Machines. Over the past few months, the one-year-old start-up has seen three co-founders and other researchers leave the company.
Andrew Tulloch, a member of technical staff who also worked at OpenAI, joined Meta in October last year, while Barret Zoph, chief technology officer, and Luke Metz, member of technical staff, rejoined OpenAI in January.
Thinking Machines will work with Nvidia to “design training and serving systems for Nvidia architectures and broaden access to frontier AI and open models for enterprises, research institutions, and the scientific community”, the company added.
First Appeared on
Source link
Leave feedback about this