Skip to content
You can now search across every topic, entity and event.What's new
CoreWeave
OrganisationUS

CoreWeave

GPU cloud provider; hosts OpenAI's compute under an $11.9bn contract.

CoreWeave, the GPU cloud operator behind OpenAI's $11.9bn compute contract, was named on 26 July as carrying a debt-to-equity ratio as high as 739 times, among the AI infrastructure lenders Apollo's chief economist called a credit-risk warning sign.

Last refreshed: 27 July 2026 · Appears in 1 active topic

Key Question

Is CoreWeave now the de facto compute backbone for OpenAI's scaled-back Stargate?

Timeline for CoreWeave

#13 11 Aug
#18 25 Jul
#3 28 Apr
View full timeline →

Background

CoreWeave is the largest independent GPU cloud operator in the United States, supplying AI training and inference infrastructure to hyperscalers and frontier AI labs unable to build fast enough on their own. It runs an $11.9 billion five-year compute contract with OpenAI signed in March 2025, and in April 2026 agreed a multi-year Claude inference deal with Anthropic. The scale of that commitment was underlined when OpenAI dropped its option on Crusoe's Abilene, Texas campus lease while deepening its CoreWeave commitment.

Founded in 2017 as the Cryptocurrency miner Atlantic Crypto by Michael Intrator, Brian Venturo, Brannin McBee and Peter Salanki, the company pivoted to GPU cloud services after the 2018 crypto crash. Nvidia invested $100 million in 2023, and CoreWeave raised $1.5 billion at a $27 billion valuation in its March 2025 IPO, one of the largest US tech listings in three years. It now operates 32 data centres across the US and Europe with roughly 250,000 GPUs.

CoreWeave's scale illustrates the hyperscaler compute bottleneck: cloud operators are years behind AI labs' appetite, making third-party GPU clouds critical swing capacity, and its rapid build-out helped drive the four major hyperscalers' record $110.75 billion combined Q1 2026 capex.

Key Issues
AI infra leverage

Its leverage draws credit-risk scrutiny

CoreWeave was named on 26 July as one of the AI infrastructure lenders whose leverage worries Apollo Global Management chief economist Torsten Slok, who compared Oracle's credit default swap spread to its 2008 crisis level and singled out CoreWeave's debt-to-equity ratio, as high as 739 times, calling the trend 'not your friend'.

The comparison matters for CoreWeave specifically because its business model depends on continuing to borrow against future GPU-lease revenue: unlike Alphabet, whose own ratio sits near 18 times, CoreWeave has no balance sheet deep enough to absorb a demand shock without refinancing risk becoming a going-concern question.

Common Questions
What is CoreWeave's contract with OpenAI?
CoreWeave signed an $11.9 billion five-year cloud compute contract with OpenAI in March 2025, making OpenAI its largest customer and supplanting Microsoft, which had accounted for over 60% of 2024 revenues.Source: Wikipedia / CoreWeave IPO filings
Did CoreWeave go public and what was its IPO price?
CoreWeave went public on Nasdaq in March 2025, raising $1.5 billion at a valuation of approximately $27 billion, making it one of the largest US tech IPOs in recent years.Source: Wikipedia
Why did OpenAI choose CoreWeave over building its own data centres?
OpenAI revised its Stargate compute strategy in 2025-26, cutting its $1.4 trillion nameplate target to roughly $600bn and preferring leased compute over owned builds. CoreWeave, with 32 ready-to-operate data centres and 250,000 GPUs, offered faster time-to-compute than greenfield construction.Source: Lowdown data-centres
How is CoreWeave different from AWS or Google Cloud?
CoreWeave focuses exclusively on GPU-accelerated workloads for AI, offering dedicated high-density compute rather than a broad cloud portfolio. This specialisation gives AI labs bare-metal GPU access without competing against general workloads for bandwidth.
Source Material