Anthropic is rapidly assembling the compute infrastructure it needs to scale its Claude AI models — and the price tag keeps climbing.
On August 31, Reuters reported that Anthropic has signed a $35 billion cloud-computing deal with Lambda, an Nvidia-backed cloud provider, for a data center in Nueces County, Texas. The facility, with a capacity of approximately 350 megawatts, is being developed by Hut 8, a crypto-mining company that has pivoted to AI data center operations.
Critically, Nvidia itself will hold the lease on the data center, according to the Wall Street Journal, which first reported the deal. The arrangement positions Nvidia not merely as a chip supplier but as an infrastructure landlord for one of the most well-funded AI companies in the world.
The Lambda deal is Anthropic's second mega-compute agreement in a single week. On August 25, the company announced it would spend $45 billion to rent AI cloud computing power from Nscale's West Virginia data center campus. Combined, the two deals total $80 billion in committed compute spending — a staggering figure even by the standards of today's AI infrastructure boom.
Anthropic, which is preparing for an IPO, has been aggressively expanding its computing capacity to meet surging demand for products like Claude Code, its AI-powered coding assistant. The company's rapid spending reflects the reality that compute access has become the primary bottleneck for AI companies seeking to train and deploy frontier models.
The Texas facility adds to a growing map of AI mega-data centers across the United States, where demand for Nvidia's advanced GPUs has outstripped supply for several consecutive quarters. By having Nvidia serve as both chip supplier and leaseholder, the Lambda deal creates a vertically integrated arrangement that ensures Anthropic gets the hardware it needs while Nvidia locks in long-term revenue from its most important customers.
Neither Anthropic, Nvidia, Lambda, nor Hut 8 immediately responded to Reuters' requests for comment.
The deal underscores a broader trend in the AI industry: the companies building the most advanced models are no longer content to rent compute on the open market. They are signing multi-decade, multi-billion-dollar commitments to secure dedicated infrastructure — a strategy that reduces supply risk but dramatically increases financial exposure.
For Anthropic, the question is whether this infrastructure investment will pay off as Claude continues to gain market share against OpenAI's ChatGPT and Google's Gemini. With over $80 billion in compute deals now on the books, the pressure to convert that capacity into revenue has never been higher.
