A montage of the CoreWeave logo and its chief executive Michael Intrator
Michael Intrator, chief executive of CoreWeave, calls his company the Tesla to Big Tech’s Ford © FT montage

An upstart challenger to cloud computing giants Amazon and Microsoft is using its alliance with Nvidia to turn artificial intelligence chips into a “new asset class” and raise billions of dollars.

New Jersey-based CoreWeave, which was founded in 2017 by former energy traders as a cryptocurrency miner, has raised $8.6bn in debt and equity over the past month to lift its valuation to $19bn. A year ago, Nvidia took a $100mn stake that valued it at $2bn.

The jump in price highlights how CoreWeave has become another beneficiary of the booming demand for AI chips that propelled Nvidia to a $2.8tn valuation and boosted the stocks of the chipmaker’s other partners such as Dell and Supermicro. Now the company is looking to put that capital to use by building facilities in the UK and Europe.

CoreWeave rents out access to its coveted stash of Nvidia chips, including the in-demand H100 and the forthcoming B200, that run inside its data centres. Its facilities were designed to accommodate the particular requirements of high-performance computing, from high-speed networking between clusters of AI chips to liquid-cooled servers, said Michael Intrator, its chief executive.

CoreWeave investors, including hedge fund Magnetar Capital, Blackstone and Coatue, are betting that soaring demand for specialised AI services will reshape the $500bn cloud computing market, despite the tens of billions that Big Tech companies are spending on their own data centres.

“The way that the cloud is going to be used for the next generation is very different than how the cloud was used in its inception 20 years ago,” said Intrator, calling the company the Tesla to Big Tech’s Ford.

Intrator said it was “incredibly hard” pitching to its earliest lenders, who had to become “experts in a field that they didn’t know anything about, to the point where they could lend billions of dollars and take it to their investment committee to propose a new asset class”, such as using Nvidia’s graphics processing units as collateral.

CoreWeave had long since pivoted away from cryptocurrencies when the release of Microsoft-backed OpenAI’s ChatGPT in November 2022 unleashed a huge wave of demand for AI computing.

Seizing its opportunity, the company quickly scaled up its financing efforts.

It raised more than $420mn in the first half of 2023 in equity, then another $2.3bn in debt financing a few months later. Some existing shareholders sold $642mn worth of stock to Fidelity and others in December. Then last month it struck two more deals to raise $7.5bn in debt and $1.1bn in equity.

CoreWeave needed the funding to “be able to scale large enough that we could support anybody who wanted to participate in the AI boom”, said Intrator, no matter how many thousands of chips they needed.

Now CoreWeave has its sights on rapid expansion in Europe. It revealed plans on Wednesday to invest $2.2bn to build three data centres in Norway, Sweden and Spain by the end of next year. It has recently committed $1.3bn for two facilities in the UK, where it is basing its European headquarters.

To expand more quickly in the US, CoreWeave this week announced a partnership with Core Scientific, a bitcoin miner, to repurpose several of its data centres to host its GPUs. Bloomberg reported on Tuesday that CoreWeave had also offered to buy Core Scientific outright for more than $1bn. The companies have not commented on the report.

Like Amazon Web Services or Microsoft’s Azure, CoreWeave offers an alternative to companies buying and maintaining their own servers, offering flexible access to computing power.

But unlike AWS, which was founded in 2006 and can host an almost endless range of applications and data, CoreWeave’s data centre serves a specific niche of clients with extremely high-performance computing demands, from AI and drug researchers to media groups.

Despite its reliance on Nvidia’s GPUs for the core of its service, Intrator argued there was a “misunderstanding” about CoreWeave’s relationship with the world’s most valuable chipmaker. “Nvidia is not giving us access to GPUs because they have some vested interest in us or because we have some advantageous access.”

Its competitive edge was more than just having the right chips, Intrator said. For instance, CoreWeave had developed software that automatically manages and maintains clusters of GPUs.

He batted off questions about whether prospective investors were concerned about backing a business that had raised capital from Nvidia, only to spend a significant portion of those funds on that company’s products.

“It’s such a crap narrative,” he said. “Nvidia invested $100mn. We’ve [raised] $12bn in debt and equity. It’s an inconsequential amount of money in the relative scale of the amount of infrastructure we’re buying.”

Nvidia also denied that companies it had invested in got preferential access to its new products. “We don’t help anybody jump the queue,” Mohamed Siddeek, head of Nvidia’s venture arm NVentures, told the Financial Times last year.

Nonetheless, Intrator said that having Nvidia examine CoreWeave’s business and agree to invest was “an incredibly powerful tool” when it came to raising capital.

“There’s an awful lot of questions that I’m able to answer based on the fact that the folks who do know more about this than anyone else are willing to put up massive amounts of capital behind what we’re doing,” he said.

He was able to convince lenders to stump up billions of dollars by leveraging a combination of its GPU assets, the value of long-term contracts signed with customers and a “proven ability to execute”, Intrator added.

By the end of 2024, CoreWeave would have 28 data centres across the US and Europe, with plans to create a “truly global footprint” over the next couple of years. “The company is continuing to scale as fast as possible,” Intrator said.

Copyright The Financial Times Limited 2024. All rights reserved.
Reuse this content (opens in new window) CommentsJump to comments section

Follow the topics in this article

Comments