CoreWeave has signed a multibillion-dollar contract with Anthropic, CEO Michael Intrator confirmed in an interview with Bloomberg Technology on April 10, 2026, capping a week in which the AI infrastructure company also announced a $21 billion agreement with Meta.
The deals represent a significant commercial milestone for CoreWeave, which went public earlier this year and has positioned itself as a critical supplier of GPU-based computing infrastructure to some of the most well-capitalised AI companies in the world. Neither CoreWeave nor Anthropic disclosed the specific financial terms of their agreement.
A Week That Reshaped CoreWeave's Commercial Standing
Intrator spoke to Caroline Hyde on Bloomberg Tech to announce the Anthropic contract, describing it as a "multibillion-dollar contract" without providing further detail on duration or scope. The agreement pairs CoreWeave with one of the AI sector's most prominent model developers — Anthropic, the company behind the Claude family of large language models, which has itself raised over $7 billion in investment and carries a valuation reported at $61 billion as of early 2025.
The Meta deal, announced separately this week, is larger and more precisely defined at $21 billion, covering the provision of computing power to support Meta's expanding AI infrastructure needs. Together, the two contracts signal that CoreWeave is consolidating relationships with the hyperscalers and frontier AI labs that are driving the bulk of current GPU demand.
CoreWeave is now locked into commercial relationships with two of the most resource-intensive AI organisations operating today — a position that could define its trajectory as a public company.
Why Infrastructure Deals at This Scale Matter
The AI industry's appetite for compute has outpaced the ability of traditional cloud providers — Amazon Web Services, Microsoft Azure, and Google Cloud — to meet specialised GPU demand on short notice. CoreWeave, which built its business around Nvidia GPU clusters, has filled that gap for companies that need dedicated, high-density compute at scale.
Anthropic's decision to contract with CoreWeave rather than rely solely on hyperscaler infrastructure reflects a broader trend among frontier AI labs: diversifying compute suppliers to secure capacity and manage costs. Training and running large language models at the scale Anthropic operates requires sustained access to thousands of GPUs over extended periods — precisely the kind of workload CoreWeave is designed to serve.
For CoreWeave, locking in long-term commitments from customers of this profile addresses one of the central concerns investors raised around its IPO earlier in 2026 — namely, whether the company's revenues were too dependent on a single major customer, Microsoft, which had accounted for a substantial share of its earlier revenue base.
CoreWeave's Path From GPU Miner to AI Infrastructure Provider
Founded in 2017 originally as a cryptocurrency mining operation, CoreWeave pivoted to AI cloud infrastructure as GPU demand accelerated. The company raised $1.1 billion in a Series C round in 2023 at a valuation of $19 billion, then filed for an IPO in early 2026. Its public debut was closely watched as a measure of investor appetite for AI infrastructure plays.
The company's business model centres on leasing Nvidia GPU capacity — primarily H100 and A100 chips — to enterprises and AI developers on a dedicated basis, rather than the shared, on-demand model offered by public cloud providers. That approach commands premium pricing but requires heavy capital expenditure upfront to acquire and operate hardware.
Securing multi-year, multibillion-dollar contracts with customers like Meta and Anthropic is therefore not just a revenue story — it is a balance sheet story. Long-term committed spend allows CoreWeave to justify continued infrastructure investment and service the debt it has taken on to build out its data centre capacity.
Anthropic's Compute Strategy Under Scrutiny
For Anthropic, the deal raises questions about the company's longer-term infrastructure strategy. Anthropic has a significant existing relationship with Google Cloud, which has invested heavily in the AI lab, and also works with Amazon Web Services through a partnership announced in 2023 in which Amazon committed up to $4 billion. Adding CoreWeave to its supplier roster suggests Anthropic is deliberately spreading compute risk across multiple vendors.
The company is in the process of training increasingly powerful iterations of its Claude models, and competition with OpenAI, Google DeepMind, and others means it cannot afford supply constraints at critical training junctures. Diversified compute contracts serve as insurance against capacity shortfalls from any single provider.
What This Means
For the AI infrastructure sector, CoreWeave's twin deals with Meta and Anthropic confirm that specialist GPU cloud providers have secured a durable role alongside hyperscalers — and that the largest AI organisations are willing to commit billions in long-term contracts to guarantee the compute access their ambitions require.