Anthropic signs $19 billion AI data center lease with TeraWulf

Artificial intelligence company Anthropic has made one of the largest long-term infrastructure commitments in the AI industry, signing a 20-year agreement worth an estimated $19 billion with AI data center operator TeraWulf to secure massive computing capacity for its rapidly expanding Claude AI platform.

Under the landmark deal, Anthropic will lease up to 401 megawatts of dedicated AI computing power at TeraWulf’s new data center campus in Hawesville, Kentucky. Initial operations are expected to begin in late 2027, with full deployment scheduled for early 2028.

The agreement reflects the enormous computing requirements needed to train and operate increasingly sophisticated artificial intelligence models. As demand for enterprise AI services continues to accelerate, major AI developers are competing to lock in long-term access to high-performance data centers capable of supporting advanced AI workloads.

Investors welcomed the announcement, sending TeraWulf shares sharply higher while boosting other AI infrastructure companies across the market. Analysts said the deal reinforces growing confidence that demand for specialized AI data centers will remain strong for years as businesses expand adoption of generative AI technologies.

Originally known for cryptocurrency mining, TeraWulf has rapidly repositioned itself as an AI infrastructure provider by converting its energy assets into high-density computing facilities. The company also announced plans to sell its stake in a Texas AI data center project to focus on wholly owned AI campuses serving large enterprise customers.

Anthropic, backed by technology giants Amazon and Google, has emerged as one of the world's leading AI companies through its Claude family of language models. As competition with OpenAI, Meta, Google, and xAI intensifies, securing reliable computing infrastructure has become a critical competitive advantage.

Industry experts say the agreement highlights a broader transformation taking place across the technology sector, where access to electricity, GPUs, cooling systems, and large-scale data centers has become just as important as advances in AI software itself.

The deal also signals continued expansion of the emerging "AI infrastructure economy," where specialized cloud providers are investing billions of dollars to build next-generation facilities designed specifically for artificial intelligence training and inference.

As AI models continue growing in complexity, long-term infrastructure partnerships like this are expected to become increasingly common, reshaping both the cloud computing industry and the future of artificial intelligence development worldwide.