Anthropic announced plans Wednesday to spend $50 billion on a U.S. artificial intelligence infrastructure build-out, starting with custom data centers in Texas and New York.
The facilities, which will be designed to support the company’s rapid enterprise growth and its long-term research agenda, will be developed in partnership with Fluidstack.
Fluidstack is an AI cloud platform that supplies large-scale graphics processing unit, or GPU, clusters to clients like Meta, Midjourney, and Mistral.
Additional sites are expected to follow, with the first locations going live in 2026. The project is expected to create 800 permanent jobs and more than 2,000 construction roles.
The investment positions Anthropic as a major domestic player in physical AI infrastructure at a moment when policymakers are increasingly focused on U.S.-based compute capacity and technological sovereignty.
“We’re getting closer to AI that can accelerate scientific discovery and help solve complex problems in ways that weren’t possible before. Realizing that potential requires infrastructure that can support continued development at the frontier,” said CEO Dario Amodei. “These sites will help us build more capable AI systems that can drive those breakthroughs, while creating American jobs.”
The move comes as Anthropic rival OpenAI pushes forward with an aggressive buildout of its own. The ChatGPT maker has secured more than $1.4 trillion in long-term infrastructure commitments through deals with Nvidia, Broadcom, Oracle, and the major cloud providers, including Microsoft, Google, and, most recently, Amazon.
The scale of that spending has raised questions about whether the U.S. has the power capacity and industrial backbone to deliver on such promises, and whether the AI sector is drifting into bubble territory.
Anthropic serves more than 300,000 businesses, with enterprise clients driving most of its revenue.
The number of large accounts, which generate over $100,000 annually, has nearly increased sevenfold in the past year. Internal projections obtained by The Wall Street Journal showed Anthropic expects to break even by 2028, well ahead of OpenAI, which is projecting $74 billion in operating losses that same year.
To support that trajectory, Anthropic tapped Fluidstack to build custom facilities optimized for its AI workloads, citing the firm’s speed and ability to deliver gigawatts of power on short timelines.
In parallel, Amazon has opened a dedicated data center campus for Anthropic on 1,200 acres in Indiana.
The $11 billion facility is already up and running, while many competitors are still promising data centers of the future. Anthropic has also expanded its compute deal with Google by tens of billions of dollars.
The move also comes as the role of the federal government in financing AI infrastructure becomes a flashpoint.
Last week, OpenAI asked the Trump administration to expand a key CHIPS Act tax credit to include AI data centers and grid components like transformers, according to a letter obtained by Bloomberg.
That request followed backlash over comments from CFO Sarah Friar, who had floated the idea of a government “backstop” for OpenAI’s compute deals.
Though the company has since walked back the suggestion of federal guarantees, the episode underscored the political and financial uncertainty surrounding how — and by whom — America’s AI infrastructure will be funded.
WATCH: SoftBank’s Nvidia exit fuels OpenAI push despite mounting losses, stiff competition from Anthropic
