According to Wall Street Journal, Anthropic has signed a cloud-computing deal worth $35 billion with Lambda, a cloud provider backed by Nvidia. The deal will bring online Nvidia capacity to meet growing demand for Anthropic's Claude AI models. As per Wall Street Journal, the most interesting aspect of the deal is Nvidia's central role behind the scenes, with the chip giant holding the lease for the data center where Anthropic's computing will take place. The facility is being built by Hut 8, a bitcoin mining company that has expanded into data centers, located in Nueces County, Texas. A few weeks before the Anthropic-Lambda agreement was finalized, Nvidia reached its own deal with Hut 8 to secure the data-center capacity. This complex three-way arrangement shows Nvidia controlling the lease, Hut 8 building and operating the physical facility, while Lambda provides the computing capacity to Anthropic.
As Anthropic scales its Claude family of AI models, the company is looking to produce hardware that can fulfill its voracious appetite for data crunching and reduce its reliance on Nvidia chips. According to Reuters, Anthropic plans to spend tens of billions of dollars to rent computing power from cloud providers and also plans to buy chips directly. The company plans to buy $36 billion worth of Google's AI chips and signed the $45 billion deal to rent AI cloud computing power from Nscale. It agreed to pay SpaceX $1.25 billion per month through May 2029 for computing capacity across its data center clusters. The discussions with MatX also suggest Anthropic may be interested in producing a training chip, while other chip startups and rival OpenAI pursue processors better suited to generating responses from chatbots.
The Lambda deal demonstrates how Nvidia is expanding its role beyond chip sales to encompass financing, data-center capacity, and partnerships that help AI firms build computing infrastructure. According to Wall Street Journal, Nvidia is holding the lease, investing in cloud providers, and providing credit support to smaller cloud companies such as Lambda in return for a share of their cloud revenue. However, Nvidia has reportedly paused parts of this programme, though it could still revamp the initiative later. Under the proposed model, Nvidia would receive 50% of any revenue cloud providers earned through its chips beyond a certain threshold. The broader Hut 8-Anthropic relationship also involves Google, with another Hut 8 facility being built nearby to house Google's competing tensor processing units (TPUs) for Anthropic, with Google providing financial guarantees to help Hut 8 raise debt for that project. Hut 8 said the facility, which is expected to cost about $20 billion, would be built to support Nvidia chips, showing how Nvidia is becoming involved in almost every part of the AI infrastructure race.
For Anthropic, which raised $8 billion from Amazon in late 2024 and early 2025, the financial resources to pursue custom silicon are clearly available. The company is investing up to $5 billion in compute partnerships, including deals with AMD, to diversify its chip supply beyond Nvidia. According to Reuters, Anthropic has hired engineering and executive talent to accelerate the chip design process. The company's IPO, expected as early as October 2026, will chase a valuation of $2 trillion, which hinges on a 2028 revenue figure of as much as $200 billion. The company's annualized revenue run rate reportedly reached about $65 billion by the end of July, compared with around $47 billion in the spring, showing rapid growth that supports its massive computing commitments. Anthropic's valuation has jumped sharply, with its Series H funding round in May valuing the company at about $965 billion, up from $380 billion just three months earlier in February.
Designing chips is expensive and time-consuming, taking a year or more to produce a viable piece of hardware, with design costs for a single generation running in the hundreds of millions of dollars. Anthropic plans to keep a multi-chip approach by working with providers from Nvidia to Google, while also planning to produce an inference chip as well. Making a custom chip may also help Anthropic hedge against the tight supply of Nvidia's processors, which Nvidia said in a conference call would be in short supply through 2027. Anthropic was one of the first companies to run its models on hardware from several vendors including Nvidia, Google and Amazon. Through this approach, the company hopes to create significant performance and economic advantages while reducing reliance on any single vendor. The shift from acquisition to partnership reflects how the biggest AI labs are thinking about hardware strategy in 2026, with MatX's founding team's TPU background giving the startup credibility that most Nvidia challengers lack, and the $500 million war chest providing runway to actually execute on a chip tape-out.
