Anthropic has reportedly signed a $35 billion cloud computing deal with Nvidia-backed Lambda for access to infrastructure at a massive new data center in Texas, according to Reuters.
The project, which is being developed by Hut 8 in Nueces County, Texas, will provide Anthropic with access to roughly 350 megawatts of capacity. Nvidia GPUs will provide the computing muscle, with the additional capacity intended to help meet growing demand for Claude.
It's an eye-watering amount of money. But what's even more remarkable is that the $35 billion agreement comes less than a week after Anthropic reportedly committed another $45 billion to rent computing power from Nscale.
In other words, Anthropic has reportedly committed around $80 billion to just these two computing deals in the span of a week.
So why does the company behind a chatbot need that much infrastructure? Let's dig into it.
Claude needs a lot more compute
If you open Claude on your phone or laptop, it's easy to forget what happens after you hit send.
Your request ultimately has to be processed by specialized hardware running inside a data center. Multiply that by millions of prompts, API calls and complicated AI agent tasks, and the infrastructure requirements quickly become enormous.
And Anthropic isn't preparing for Claude's current level of demand. It's building for where it thinks that demand is headed.
Anthropic said in May that its annualized revenue had surpassed $47 billion, up from approximately $9 billion at the end of 2025. The company specifically said some of the money from its latest funding round would be used to "expand compute to meet growing demand for Claude."
Claude Code appears to be particularly important to growth
Anthropic said earlier this year that Claude Code's annualized revenue had climbed above $2.5 billion and that weekly active users had doubled since the beginning of 2026. Business subscriptions to the coding agent had quadrupled over the same period.
That's important because tools such as Claude Code can require considerably more interaction with an AI model than asking a chatbot a single question.
Instead of producing one response and stopping, an AI coding agent can inspect files, write code, run commands, analyze the results, fix mistakes and continue working through a task. Anthropic's own research into roughly 400,000 Claude Code sessions found that usage has been shifting toward more end-to-end agentic tasks, including running software, analyzing data and producing documents.
The more work we hand over to AI agents, the more inference compute companies such as Anthropic need behind them.
$35 billion isn't even Anthropic's biggest recent compute deal
Just days ago, Reuters reported that Anthropic had agreed to spend $45 billion over six years for computing capacity from Nscale's data center campus in West Virginia.
That agreement covers about 460 megawatts of capacity and is expected to use Nvidia's next-generation Vera Rubin hardware.
Anthropic has been assembling compute capacity from multiple directions. It announced a $50 billion U.S. data center investment with Fluidstack last year, and in May said a partnership with SpaceX would substantially increase the amount of compute available for Claude.
That SpaceX capacity had an immediate impact users could actually see: Anthropic doubled Claude Code's five-hour rate limits for Pro, Max, Team and some Enterprise users, removed a peak-hours restriction and raised API limits for Claude Opus models.
It's a useful example of why these enormous infrastructure deals ultimately matter to people using Claude. More compute doesn't automatically mean a smarter AI model, but it can determine how many people can use it, how frequently they can use it and how much work increasingly autonomous AI agents can perform.
Nvidia is everywhere in this deal
There's another interesting wrinkle to Anthropic's latest agreement: Nvidia seems to be everywhere.
Lambda is backed by Nvidia. The data center will contain Nvidia hardware. And according to The Wall Street Journal, Nvidia itself will hold the lease on the Texas data center.
Nvidia is also an Anthropic investor. The companies announced a broader strategic partnership last November that included plans for Anthropic to use Nvidia hardware and for the two companies to work together on optimizing future Nvidia architectures for Anthropic's AI workloads.
Even Anthropic's $45 billion Nscale agreement reportedly involves Nvidia's upcoming Vera Rubin chips.
That shows just how central Nvidia has become to the infrastructure sitting behind the current AI boom. Companies such as Anthropic may be competing with OpenAI, Google and others to build the best AI models, but much of that race still depends on securing enormous quantities of computing hardware and electricity to run them.
What this means for Claude users
Don't expect your Claude subscription to suddenly look different because Anthropic signed another data center deal. This is a much longer-term bet.
Anthropic is effectively wagering tens of billions of dollars that Claude usage — particularly through products such as Claude Code and future AI agents — is going to continue growing dramatically.
We've already seen what additional computing capacity can do. Anthropic used its SpaceX deal to raise Claude usage limits earlier this year. More infrastructure could similarly help the company accommodate more users, longer-running agents and increasingly compute-intensive models without constantly running into capacity constraints.
And with another $35 billion reportedly going toward Claude's computing needs, Anthropic clearly doesn't expect our appetite for AI to slow down anytime soon.
Follow Tom's Guide on Google News and add us as a preferred source to get our up-to-date news, analysis, and reviews in your feeds. Subscribe to Tom's Guide on YouTube and follow us on TikTok.
More from Tom's Guide
