On November 12, 2025, Anthropic announced on its official website that it will invest $50 billion in AI computing infrastructure in the United States, build a new generation of data centers in Texas and New York with partner Fluidstack, and plan to select sites in more regions in the future. Officials said that these data centers are custom-designed for Anthropic's workloads and are designed to maximize computing power efficiency to support the continuous research and development of cutting-edge large models.
According to the company's disclosure, the project is expected to create about 800 long-term jobs and 2,400 construction-phase jobs, and the data center will be online from 2026. Anthropic linked the investment to the U.S. federal-level AI Action Plan, emphasizing that the move helps solidify the U.S. leadership in AI and strengthen its indigenous technology infrastructure. CEO Dario Amodei said that a more robust infrastructure is a prerequisite for driving scientific discovery and solving complex problem-solving AI applications.
Anthropic also mentioned in the announcement that its growth is driven by high-density technical teams, security and alignment research, and enterprise application expansion around the Claude family of models, currently serving more than 300,000 enterprise customers, and the number of large accounts has increased nearly sevenfold in the past year. Fluidstack was chosen as a partner for its ability to quickly deliver gigawatt-scale power and large-scale GPU clusters. The company also emphasized that it will prioritize cost-effectiveness and capital efficiency while pursuing scale to cope with the sharp rise in computing power and energy demand across the industry.
Frequently Asked Questions
Q: What projects is the $50 billion invested in?
A: Anthropic said the funds will be mainly used to build custom AI data centers in the United States, including new facilities in Texas and New York, as well as other sites that may be expanded in the future to deploy large-scale computing power clusters and related supporting infrastructure.
Q: When will these data centers start operating?
A: According to official information, the first batch of sites is expected to be launched in 2026, and a large number of construction jobs will be brought during the construction period, providing long-term operation and maintenance and technology-related positions after production.
Q: How does this investment relate to the U.S. government's AI Action Plan?
A: Anthropic clearly mentioned in the announcement that the project will help implement the goals of the AI Action Plan proposed at the federal level, including maintaining the United States' leadership in artificial intelligence and strengthening local technology infrastructure and employment capabilities, so the company positions it as a commitment to "American AI leadership".
Q: Why did Anthropic shift from "renting clouds" to building its own data centers?
A: From the perspective of public information and industry analysis, with the sharp rise in the demand for computing power, bandwidth and energy by large models such as Claude, it is difficult to maintain an advantage in cost and resource acquisition by relying only on the public cloud.
Q: What does this mean for other AI and cloud computing companies?
A: Anthropic's $50 billion plan is seen as part of the "arms race" for AI infrastructure in the United States, and the huge investment in computing power and data centers by other technology companies means that top AI labs are shifting from software-only companies to asset-heavy infrastructure players. This will intensify competition for power, land, chips and professionals, and will also amplify regulatory and social discussions around energy consumption and environmental impact.
