(Reuters) -AI startup Anthropic said on Wednesday it would invest $50 billion in building data centers in the U.S., the latest multi-billion-dollar outlay in the industry as companies race to expand their artificial intelligence infrastructure.

The company behind Claude AI models said it would set up the facilities with infrastructure provider Fluidstack in Texas and New York, with more sites coming online in the future. The data centers are custom-built for Anthropic.

Tech companies have announced massive spending plans this year, with many focusing on expanding their U.S. footprint, as President Donald Trump pushes for investments on American soil to maintain the country's edge in the AI sector.

Trump ordered his administration in January to produce an AI Action Plan that would make "America the world capital in artificial intelligence". As part of the push, several American companies had rolled out a series of big-ticket AI and energy investment pledges at Trump's tech and AI summit in July.

Anthropic said the project is expected to create about 800 permanent jobs and 2,400 construction jobs in the U.S. as the data centers come online throughout 2026.

The outlay "will help advance the goals in the Trump administration's AI Action Plan to maintain American AI leadership and strengthen domestic technology infrastructure", Anthropic said.

The San Francisco-based company, which is backed by Google-parent Alphabet and Amazon.com ( AMZN ), was valued at $183 billion in early September.

Formed in 2021 by a group of former OpenAI employees, Anthropic serves more than 300,000 enterprise customers. Its Claude large language models are widely regarded as one of the most powerful frontier models on the market.

(Reporting by Deborah Sophia in Bengaluru; Editing by Arun Koyyur)