Meta and CoreWeave expand AI cloud deal to $21 billion through 2032

Back to home

Share

M&A

1 min read

|

Updated 2 months ago

|

$META $CRWV $NVDA

Briefing

Meta Platforms and CoreWeave have extended their AI cloud infrastructure partnership in a deal valued at approximately $21 billion, running through December 2032. The agreement grants Meta access to CoreWeave's AI cloud capacity, including Nvidia's next-generation superchips. Both companies' shares rose on the news.

Meta and CoreWeave lock in $21 billion AI compute deal to 2032

Meta Platforms and CoreWeave have expanded their existing AI cloud partnership into a long-term contract worth approximately $21 billion, according to the Wall Street Journal, which first reported the terms. The agreement runs through December 2032 and provides Meta with AI cloud infrastructure capacity from CoreWeave, the GPU-focused cloud provider that listed on Nasdaq earlier this year.

The deal encompasses access to Nvidia's next-generation superchips, according to CoinDesk, reflecting Meta's stated ambition to scale both AI training and inference workloads well into the next decade. The duration and scale of the contract are notable: a seven-year commitment of this size signals that Meta is treating third-party cloud capacity as a structural rather than supplemental component of its AI infrastructure strategy.

Both stocks gained on Thursday following the announcement, according to Decrypt, though specific price moves were not disclosed in the source reporting.

For CoreWeave, the contract represents a significant revenue anchor from one of the world's largest AI spenders. Meta has previously indicated it plans to invest heavily in AI infrastructure and has been one of the most active builders of large language models outside the dedicated AI lab sector.

Analysis

A seven-year, $21 billion contract converts CoreWeave from a growth-stage GPU leasing story into a contracted-revenue infrastructure business. This reduces CRWV's revenue concentration risk perception but also caps the upside narrative: Meta's commitment implies CoreWeave negotiated at scale pricing, compressing the per-GPU margin expansion thesis that justified post-IPO growth multiples. Investors in CRWV must now reprice from venture-style optionality to utility-style visibility.
Meta committing to third-party GPU cloud capacity through 2032 alongside its own capex buildout signals that even the largest hyperscale AI spenders cannot self-provision fast enough to meet training and inference demand timelines. This structurally validates the merchant cloud GPU model and raises the probability that other frontier AI labs and large-cap tech companies sign comparable long-duration contracts with CoreWeave or its direct competitors, tightening available GPU allocation for smaller customers.
The deal is denominated in Nvidia next-generation superchip access, meaning NVDA sits upstream of the entire $21 billion commitment. CoreWeave's contracted revenue is only deliverable if Nvidia hits production and allocation targets, embedding NVDA supply execution risk into Meta's AI roadmap through 2032. Any Nvidia supply disruption or export control action affecting next-generation chips now carries a direct P&L consequence for both Meta and CoreWeave simultaneously.

Connected Events

Markets · 2 months ago

Meta launches Muse Spark LLM in direct challenge to OpenAI and Google

Meta's simultaneous launch of Muse Spark LLM positions the $21 billion CoreWeave deal as direct infrastructure support for a live frontier model product, not speculative future capacity. The training and inference workloads for a model being integrated into Instagram and Facebook at scale require exactly the kind of long-duration, high-volume GPU access this contract provides, compressing the timeline between capital commitment and revenue-generating deployment.

Markets · 2 months ago

OpenAI halts Stargate UK data centre project over energy costs and regulation

OpenAI's suspension of its Stargate UK data centre project over energy costs and regulatory friction reinforces the competitive advantage of CoreWeave's US-based GPU capacity. As European AI infrastructure buildout stalls, demand concentration in US merchant GPU cloud providers increases, improving CoreWeave's negotiating position with future customers and validating Meta's decision to lock in long-duration US capacity now.

Meta and CoreWeave lock in $21 billion AI compute deal to 2032

Analysis

Connected Events

Meta launches Muse Spark LLM in direct challenge to OpenAI and Google

OpenAI halts Stargate UK data centre project over energy costs and regulation

Further reading

Related Stories

Innovent Biologics signs $10.5bn oncology collaboration with Pfizer

Pentagon awards Dell $9.7bn software contract to consolidate licenses

Meta and CoreWeave expand AI cloud deal to $21 billion through 2032

Anthropic raises $65bn Series H at $965bn valuation, eclipsing OpenAI