Anthropic Signs $35 Billion Cloud Deal With Nvidia-Backed Lambda

Anthropic has signed a $35 billion cloud computing deal with Lambda, a cloud provider backed by Nvidia, to secure compute for a Texas data center being developed by Hut 8 in Nueces County. The facility covers roughly 350 megawatts of capacity, with Nvidia holding the lease on the data center itself and installing its own chips there — an arrangement that let Lambda, which counts Nvidia as an investor, land a major Anthropic contract without having to independently source or lease the underlying real estate.

One Deal in a Fast-Moving Sequence

This Lambda agreement is not an isolated event — it lands within weeks of Anthropic committing $45 billion to rent capacity from Nscale's West Virginia data center campus, and a separate $50 billion deal with neocloud provider Fluidstack. Taken together, Anthropic has committed well over $100 billion in compute contracts in a single stretch of weeks, aimed squarely at keeping pace with demand for Claude.

The Structural Pattern Behind These Deals

• Nvidia is increasingly the structural anchor of these arrangements, not just the chip supplier — holding data center leases directly and enabling smaller neoclouds like Lambda to win large AI-lab contracts they couldn't otherwise finance alone
• Each of Anthropic's recent mega-deals (Lambda, Nscale, Fluidstack) is with a different specialized 'neocloud' provider rather than a single hyperscaler, spreading both capacity risk and counterparty concentration across multiple vendors
• The pace of commitment — three deals worth over $100 billion combined within weeks of each other — reflects compute scarcity and demand growth that outstrips what any single provider can supply alone

What This Means If You Build on Claude or Evaluate AI Vendors

For enterprises building on Claude, this scale of committed capacity is a reasonable signal against near-term compute-driven outages or throttling, though it says nothing directly about pricing. For anyone evaluating AI vendor stability more broadly, the more useful habit is tracking total committed spend against disclosed revenue over time, not treating any single headline deal in isolation — the pattern of financing (debt, neocloud partnerships, equity stakes) matters as much as the dollar figure itself.