Nvidia expands AI influence with Anthropic’s $35B Lambda partnership deal

Anthropic has struck a cloud-computing deal valued at approximately $35 billion with Lambda, an Nvidia-backed cloud provider, highlighting Nvidia’s rapidly expanding influence in the artificial intelligence infrastructure ecosystem. The arrangement, detailed by the Wall Street Journal, is notable not just for its size but also for the multi-layered role Nvidia plays, as it continues to move far beyond being just a supplier of advanced AI chips.

In this agreement, Nvidia has secured the lease for the data centre in Nueces County, Texas, where Anthropic’s computing will take place. This facility is being built and operated by Hut 8, a bitcoin-mining company now pivoting into large-scale data centre operations. While Lambda acts as the intermediary providing computing capacity to Anthropic using Nvidia’s chips, actual control over the facility rests with Nvidia as the leaseholder, and the chips also come from Nvidia, which is a Lambda investor. The data centre, expected to cost around $20 billion, will be tailored specifically for Nvidia hardware, according to Hut 8.

The partnership exemplifies how Nvidia’s role in the AI arms race now covers financing, securing data-centre capacity, and structuring partnerships that help AI firms gain critical computing power. The precise financial details between Lambda and Nvidia, including lease terms and revenue-sharing arrangements, remain undisclosed.

Anthropic’s broader data centre expansion goes well beyond the Lambda deal. In July, it signed a $45 billion, six-year commitment with Nscale, another Nvidia-backed provider, for 460 megawatts of capacity at a West Virginia data centre that will run Nvidia’s next-generation Vera Rubin chips starting in late 2027. Anthropic’s contract is central for Nscale, enabling the startup to use the agreement as collateral to secure funding for the yet-to-be-built site. In total, just the Lambda and Nscale pacts are worth about $80 billion.

Anthropic has also made significant deals with Fluidstack ($50 billion), Norway-based Volta ($10 billion), and AMD, with up to $5 billion in investment from AMD tied to the deployment of its Instinct MI450 chips by 2027. Another major agreement with SpaceX provides approximately $1.25 billion per month in computing resources from two SpaceX-operated sites. Meanwhile, Amazon has expanded its partnership with Anthropic to include multiple gigawatts of capacity via its cloud and Trainium chips, committing up to $33 billion in capital, while Google and Microsoft have also ramped up investments. Overall disclosed computing commitments by Anthropic now top $150 billion.

These agreements come as Anthropic reportedly prepares for a public offering as soon as October, having confidentially filed with the US SEC in June 2026. Goldman Sachs, JPMorgan, and Morgan Stanley are expected to run the IPO, which could see the company seek a public valuation as high as $2 trillion, according to some reports. Anthropic’s rapid growth, with annualised revenue reaching about $65 billion by July, is central to its investment case as it races to meet exploding demand for its Claude AI models and maintain its competitive edge against industry rivals like OpenAI.