In brief: Meta Platforms has announced an additional investment of $21 billion in CoreWeave, aimed at enhancing AI cloud capacity from 2027 through December 2032. This brings the total investment between the two companies to about $35 billion. The new contract will prioritize early deployments of Nvidia’s Vera Rubin platform tailored for inference workloads, rather than training. CoreWeave will also pursue $4.25 billion in new financing to fuel its expansion, including $3 billion in convertible notes and $1.25 billion in junk bonds. Following the news, CoreWeave shares experienced a 5% increase, while Meta's stock rose by approximately 3%.
From Cryptocurrency Mining to AI Cloud Giant
CoreWeave, originally founded in 2017 as Atlantic Crypto for Ethereum mining, pivoted to GPU cloud infrastructure after the 2018 cryptocurrency crash. With Ethereum's shift to proof-of-stake, the founders—Michael Intrator, Brian Venturo, and Brannin McBee—recognized the potential for their GPU assets in machine learning. Renamed CoreWeave in 2019, the company went public on March 28, 2025, at a share price of $40, valuing it at $23 billion. Its revenue surged to $5.13 billion in 2025, reflecting a 168% year-over-year increase, with a contracted backlog exceeding $66 billion. The first agreement with Meta, worth $14.2 billion, was announced in September 2025, establishing CoreWeave as a competitor to larger hyperscale cloud providers. This latest expansion on April 9, 2026, further cements Meta's role as CoreWeave's key commercial partner, ensuring financial stability into the next decade.
Understanding Meta's Acquisition
The contract focuses on inference rather than training, as Meta's Llama model family is open-weight and already trained before any cloud contract is executed. The ongoing operational costs involve serving these models to billions of users on platforms like Facebook, Instagram, and WhatsApp. At Meta's scale, low-latency computing across distributed infrastructures is crucial, especially during peak usage times when their data centers may be overwhelmed. CoreWeave's deployment will include some of the first commercial applications of Nvidia’s Vera Rubin platform, introduced at GTC 2026. This agreement complements Meta’s ongoing infrastructure development, with planned capital expenditures of $115 to $135 billion in 2026, primarily driven by AI infrastructure needs. Earlier in March 2026, Meta signed a $27 billion deal with Nebius for dedicated compute services, also incorporating Vera Rubin technology, emphasizing a strategy to diversify its cloud infrastructure.
Diversifying Customer Base
For CoreWeave, the $21 billion commitment from Meta helps mitigate revenue concentration risks that have plagued the company since its IPO. Previously, Microsoft accounted for 62% of CoreWeave's 2024 revenue, a scenario that raised concerns among institutional investors. With the new Meta deal, CoreWeave CEO Michael Intrator stated that no single customer will now exceed 35% of total sales, minimizing risk associated with dependence on one major client. Nvidia remains central to CoreWeave's operations, with the company having built its entire infrastructure around Nvidia GPUs. The new Meta contract extends this partnership into the next generation of hardware. Additionally, CoreWeave has enhanced its agreement with OpenAI by up to $6.5 billion, further broadening its clientele beyond Microsoft. Following the announcement of the Meta expansion, CoreWeave's stock fluctuated between $88 and $95, up from a peak of $187 in mid-2025.
Funding the Future
Building AI cloud infrastructure involves significant upfront costs, and CoreWeave has largely financed its growth through debt. In conjunction with the Meta deal announcement, the company revealed plans to raise $4.25 billion through $3 billion in convertible senior notes due by 2032 and $1.25 billion in senior unsecured notes due by 2031. CoreWeave's total debt has surged to approximately $30 billion, tripling in just a year. The company argues that its substantial backlog of over $66 billion in contracted revenue provides ample coverage for its debt obligations. Intrator has likened CoreWeave to an “AI factory,” where long-term customer commitments help underwrite capital expenses before infrastructure is established. The financing landscape for AI infrastructure has evolved dramatically, requiring large-scale debt instruments that were not available two years ago. CoreWeave has positioned itself as a vital player in the AI infrastructure sector, securing commitments that underscore its growth trajectory.