Microsoft ink handles CoreWeave to meet OpenAI cloud needs

Microsoft’s massive investment in OpenAI has put the company at the center of the artificial intelligence boom. But that’s not the only place where the software giant is opening its wallet to meet the growing demand for AI-powered services.

CNBC has learned from people with knowledge of the matter that Microsoft has agreed to spend potentially billions of dollars over several years on cloud infrastructure from startup CoreWeave, which on Wednesday announced it raised $200 million. That funding comes just over a month after the company hit a $2 billion valuation.

CoreWeave sells simplified access to Nvidia’s graphics processing units, or GPUs, which are considered the best available on the market for running AI models. Microsoft signed the CoreWeave deal earlier this year to ensure that OpenAI, which powers the viral ChatGPT chatbot, will have sufficient computing power going forward, said one of the people, who asked not to be named due to confidentiality. OpenAI relies on Microsoft’s Azure cloud infrastructure for its big data needs.

Microsoft and CoreWeave both declined to comment.

The generative AI rush began late last year after OpenAI introduced ChatGPT to the public, demonstrating that AI can take human input and produce sophisticated responses. Many companies, including Google, have since rushed to add generative artificial intelligence to their products. And Microsoft has been busy releasing chatbots for its own services, such as Bing and Windows.

With so much demand on the infrastructure, Microsoft needs more ways to tap Nvidia’s GPUs. CoreWeave CEO Michael Intrator declined to comment on the Microsoft deal in an interview last month, but he said revenue has “increased by many multiples from 2022 to 2023.”

CoreWeaves’ announced funding on Wednesday from hedge fund Magnetar Capital was an extension of a $221 million round in April. Nvidia invested $100 million in the earlier funding, Intrator said. CoreWeave was founded in 2017 and has 160 employees.

Nvidia’s share price is up 170% this year. The company’s market capitalization briefly topped $1 trillion for the first time this week after it issued a forecast for the July quarter that was more than 50% higher than Wall Street’s estimate.

The chipmaker’s growth will be “largely driven by data centers, reflecting a sharp increase in demand related to generative AI and large language models,” Colette Kress, Nvidia’s chief financial officer, said on last week’s earnings call. OpenAI’s GPT-4 large language model, trained with Nvidia GPUs on extensive online data, is at the heart of ChatGPT.

Kress referred to CoreWeave by name on the call, and in March Nvidia CEO Jensen Huang mentioned CoreWeave in his presentation at Nvidia’s GTC conference.

CoreWeave’s website claims the company can deliver computing power that is “80% cheaper than legacy cloud providers.” CoreWeave offers, among other things, Nvidia’s A100 GPUs, which developers can also find through the Amazon, Google and Microsoft clouds.

In addition, CoreWeave has available more affordable Nvidia A40 GPUs marketed for visual computing, while the A100 targets AI, data analytics and high-performance computing. Some CoreWeave clients have struggled to get enough GPU power on large clouds, Intrator said. Sometimes prospects have asked for A100 or newer H100 GPUs from Nvidia, and the company has recommended A40 GPUs instead.

These “will do an excellent job at a very cost-effective price,” Intrator said.

Microsoft has been in discussions with Oracle about the two companies renting servers from each other if they need extra capacity, The Information reported earlier this month, citing an unnamed person.

SEE: Microsoft is “just getting started” in the AI ​​space, says DA Davidson’s Gil Luria