Anthropic has agreed to rent data center capacity from CoreWeave to support growing demand for its Claude AI models, according to a report by Bloomberg Technology published on April 10, 2026.
The deal adds Anthropic to the growing list of major AI companies turning to CoreWeave — a GPU-focused cloud provider that completed a high-profile IPO in March 2025 — for computing infrastructure. CoreWeave has positioned itself as a preferred alternative to hyperscalers like AWS and Microsoft Azure for AI workloads, offering access to large clusters of Nvidia GPUs purpose-built for model training and inference.
Why Anthropic Is Betting on CoreWeave
Founded in 2021 by former OpenAI researchers, Anthropic has raised more than $12 billion to date, including a landmark $4 billion investment from Amazon and a further $2.75 billion from Google. Despite access to Amazon Web Services infrastructure through that partnership, the CoreWeave agreement signals that Anthropic is diversifying its compute supply chain to avoid bottlenecks as user demand for Claude scales.
Anthropic is diversifying its compute supply chain to avoid bottlenecks as user demand for Claude scales.
Compute availability has become one of the defining competitive constraints in the AI industry. Training and serving large language models requires tens of thousands of GPUs operating in coordinated clusters, and demand consistently outpaces supply. By securing dedicated capacity from CoreWeave, Anthropic gains a more predictable infrastructure layer — critical as it pushes Claude into enterprise contracts that require reliability guarantees.
CoreWeave's Growing Roster of AI Clients
CoreWeave went public on the Nasdaq in March 2025, raising approximately $1.5 billion at a valuation of around $23 billion. The company has since pursued major commercial agreements as proof of sustained revenue. Its existing clients include Microsoft, which signed a $10 billion multi-year deal, and OpenAI, which agreed to an $11.9 billion contract. Adding Anthropic reinforces CoreWeave's position as a neutral infrastructure provider to competing AI labs — a strategically valuable role that insulates the company from picking winners in the model race.
Financial terms of the Anthropic deal were not disclosed, according to Bloomberg. However, given the scale of comparable agreements CoreWeave has announced, industry analysts expect multi-year commitments from major AI clients to run into the hundreds of millions or billions of dollars.
Claude's Expanding Commercial Footprint
Anthropic has been accelerating its commercial push throughout 2025 and into 2026. The company launched Claude for Enterprise, expanded its API developer platform, and integrated Claude models into third-party products across coding, customer service, and document analysis. Each new customer segment adds inference demand — the compute required to generate responses in real time — making reliable, scalable GPU access a direct business priority rather than a background operational concern.
The agreement also comes as Anthropic faces intensifying competition. OpenAI released GPT-4o and a series of reasoning-focused models, while Google DeepMind has pushed Gemini aggressively into enterprise channels. To compete effectively, Anthropic must ensure Claude remains fast and available even during demand spikes — something that requires robust infrastructure commitments made well in advance.
What the Deal Reveals About AI's Infrastructure Economy
The broader pattern here is instructive. AI labs, regardless of their cloud partnerships with hyperscalers, are increasingly signing direct compute agreements with specialist providers like CoreWeave. This reflects both the scarcity of high-density GPU clusters and a desire for infrastructure independence. Anthropic's Amazon relationship does not appear to be an obstacle to sourcing additional capacity elsewhere — a sign that the compute market remains fragmented enough that no single provider can meet every lab's needs.
CoreWeave, for its part, benefits from locking in long-term revenue from creditworthy clients whose demand is structurally tied to AI adoption rather than cyclical IT spending. The more Claude is used, the more Anthropic pays — a durable revenue dynamic for any infrastructure provider.
What This Means
For businesses and developers building on Claude, this deal reduces the risk of capacity-driven service degradation as Anthropic scales — and signals that the company is investing seriously in the infrastructure required to compete at the frontier of AI.