Anthropic has signed a $35 billion cloud computing agreement with Lambda, the Nvidia-backed AI cloud provider, according to an exclusive report from The Wall Street Journal that was quickly confirmed by Reuters, Bloomberg, and The Information. The deal, one of the largest compute commitments disclosed by any AI lab to date, hands the maker of Claude a vast pool of GPU capacity as the race to train and serve frontier models increasingly resembles an arms race in infrastructure.
The agreement was first reported on August 31, 2026, with Reuters citing a source familiar with the matter. Neither Anthropic nor Lambda has published the full terms, and details such as the contract's duration and the specific data centers involved have not been disclosed. For more context on this story, see our ongoing AI news.
A Landmark Compute Agreement
The headline number is striking even by the standards of today's AI industry, where nine-figure infrastructure deals have become routine and ten-figure commitments are no longer rare. A $35 billion cloud computing contract places Anthropic in the company of the largest enterprise technology buyers in the world, and it does so not with a hyperscaler but with one of the leading so-called neoclouds: specialist GPU cloud providers that have emerged to serve AI companies' voracious appetite for compute.
According to the initial reporting, the deal centers on cloud computing capacity — the GPU-backed infrastructure used both to train large models and to serve hundreds of millions of inference requests. Reuters, the Wall Street Journal, Bloomberg, and The Information all reported the same topline figure within hours of each other, lending the story unusually broad confirmation for an unnamed-source report.
Who Is Lambda?
Lambda is one of the most prominent of the new generation of AI-focused cloud providers. The company operates data centers purpose-built for GPU workloads and counts Nvidia among its backers, a relationship that gives it privileged access to the chipmaker's most sought-after hardware in a market where supply remains constrained.
The company made headlines late last month when it secured $1 billion in private debt financing to purchase Nvidia chips that it leases out to Microsoft, as AI Buzz Wire reported at the time. That arrangement made Microsoft one of Lambda's marquee customers; the new Anthropic deal adds a second frontier AI lab to that roster and signals that Lambda's model of financing, building, and renting out AI infrastructure has moved from promising to proven.
For Lambda, the Anthropic contract provides the kind of long-duration, investment-grade revenue that underpins billions of dollars in data center construction. For Anthropic, it provides dedicated capacity at a moment when access to compute has become the binding constraint on frontier AI development.
Anthropic's Widening Infrastructure Race
The Lambda agreement is the latest in a series of aggressive infrastructure moves by the San Francisco-based lab. In late July, banks lined up a $15 billion loan for an Anthropic AI data center in Texas, a financing package backed by Google, which is both an investor in the company and a cloud partner. Anthropic has also built an in-house chip design team as it explores custom silicon to power future versions of Claude, hiring former Google TPU architects to accelerate that effort.
The spending has been matched by financial ambition. Anthropic is reportedly preparing for an initial public offering that could rank among the largest technology debuts in history, with press reports suggesting the company could seek to raise as much as $100 billion and internal revenue forecasts said to approach $200 billion by 2028. The company's quarterly profit is reported to be approaching $1 billion as enterprise demand for Claude accelerates.
Against that backdrop, the Lambda deal reads as both a supply commitment and a statement: Anthropic intends to control its compute destiny, diversifying beyond its hyperscaler partnerships and locking in capacity before competitors can claim it.
Why Compute Deals Define the Frontier AI Era
Compute has become the industry's scarcest resource. Training runs for frontier models now require clusters of hundreds of thousands of GPUs operating for months, and serving the resulting models to hundreds of millions of users multiplies the demand further. Every major AI lab has responded the same way: signing multi-billion-dollar agreements with cloud providers, chipmakers, and data center operators years in advance.
The result is an infrastructure boom that is reshaping capital markets, energy grids, and the semiconductor supply chain. Nvidia, whose GPUs power the vast majority of AI training, has seen its valuation swell to historic heights, and the companies that can broker access to those chips — Lambda chief among them — have become strategic chokepoints in the AI economy.
The Lambda deal also underscores a subtle shift in the market's structure. A year ago, frontier labs concentrated their spending with the three hyperscalers. Now, labs are spreading commitments across specialist providers that can move faster and dedicate entire facilities to a single customer. That diversification gives labs negotiating leverage and resilience, but it also fragments the supply chain in ways that analysts warn could amplify the impact of any future demand shock.
What Comes Next
Neither company has said when the capacity covered by the deal will come online, how it will be split between training and inference, or whether the agreement includes options for expansion. Those details will matter for investors poring over Anthropic's reported IPO preparation, since compute commitments of this size flow directly through to capital expenditures and operating costs.
What is already clear is the direction of travel. The frontier AI race is now as much about securing electricity, land, chips, and financing as it is about research talent. With $35 billion committed to a single supplier, Anthropic has placed one of the industry's clearest bets yet that compute, not code, will decide who leads the next phase of artificial intelligence.
Both Anthropic and Lambda declined to comment beyond the reported terms, according to the outlets that broke the story. This article will be updated as more details of the agreement are confirmed.
---
Stay Ahead of AIGet the latest AI news, analysis, and breakthroughs — all in one place.
Read more AI news →
