Michael Intrator, Chief Executive Officer of CoreWeave Inc., speaks during an interview with CNBC on the floor at the New York Stock Exchange (NYSE) in New York City, U.S., Sept. 22, 2025.
Jeenah Moon | Reuters
Shares of CoreWeave closed up nearly 12% on Tuesday after the company announced it has agreed to provide Meta with $14.2 billion of artificial intelligence cloud infrastructure.
CoreWeave has been on a deal-making blitz as big tech companies and AI startups race to build out their computing infrastructure. Tuesday’s announcement comes just days after CoreWeave expanded its agreement with OpenAI by $6.5 billion, bringing the total contract to $22.4 billion.
“The agreement underscores that behind every AI breakthrough are the partnerships that make it possible,” a CoreWeave spokesperson said in a statement about the Meta deal on Tuesday.
Meta did not immediately respond to CNBC’s request for comment. Bloomberg was first to report the deal.
CoreWeave primarily generates revenue by building and renting out data centers that are full of Nvidia‘s graphics processing units, which are key for training models and running large AI workloads. The company, which some investors have classified as a “neocloud,” has become a crucial player in an increasingly interconnected web of AI infrastructure partners.
Meta has the option to “materially expand its commitment” for additional computing capacity through 2032, according to a filing with the U.S. Securities and Exchange Commission on Tuesday.
The social media company has been pouring tens of billions of dollars into AI in recent months in an effort to build out its AI infrastructure and bring in top talent.
Meta said during its second-quarter earnings call in July that it expects its total expenses for 2025 to fall between $114 billion and $118 billion, and the company expects its AI initiatives will “result in a 2026 year-over-year expense growth rate that is above the 2025 expense growth.”
In July, Meta CEO Mark Zuckerberg said Meta is on track to bring its first supercluster online next year, and it is building several others. A supercluster is a large, complex computing network that’s housed within a data center.
One of those facilities, which Meta has named Hyperion, is expected to be large enough to cover “a significant part of the footprint of Manhattan,” Zuckerberg said.
“Meta Superintelligence Labs will have industry-leading levels of compute and by far the greatest compute per researcher,” Zuckerberg wrote in a Facebook post.
WATCH: CoreWeave CEO: Building AI infrastructure will require trillions in public-private investment
https://www.cnbc.com/2025/09/30/coreweave-meta-deal-ai.html
