en.Wedoany.com Reported - U.S. data center developer and operator TeraWulf has signed a 20-year lease agreement with artificial intelligence company Anthropic, expected to generate $19 billion in contract revenue during the initial phase. TeraWulf will build a dedicated AI infrastructure campus within its Justified Data complex in Horseville, Kentucky, featuring approximately 401 megawatts of critical IT load. The campus will be developed in phases, with initial capacity expected to come online in the second half of 2027 and full operational capability achieved by early 2028, as confirmed by TeraWulf in its official announcement.
TeraWulf Chairman and CEO Paul Prager stated that the company had previously signaled to investors an anticipated major commitment from a significant customer around the end of the second quarter. The contract with Anthropic validates the company's business strategy and establishes a long-term revenue stream with one of the world's leading AI companies. He added that the agreement lays the foundation for future expansion and demonstrates TeraWulf's ability to provide energy, develop infrastructure, and secure long-term customer commitments.
Meanwhile, TeraWulf announced a separate agreement to sell its entire 50.1% stake in a joint venture with AI cloud platform Fluidstack, which is building a 168-megawatt data center campus in Abernathy, Texas. The acquisition is to be completed by an investor group led by Fluidstack, and TeraWulf will realize approximately $450 million in investment through this transaction at a premium over invested capital. Following the transaction, Fluidstack will continue to lead the project.
This contract underscores that power supply has become critical to the expansion of generative AI. AI campuses are no longer planned solely based on land availability and fiber connectivity, but primarily on the ability to secure hundreds of megawatts of stable power over the long term. A 401-megawatt load is equivalent to the consumption of a large industrial complex, requiring years of grid interconnection planning.
Such contracts provide infrastructure providers with long-term revenue predictability but also expose them to risks from rising energy, equipment, and financing costs. For Anthropic, this lease is part of a broader strategy to secure the computing capacity needed to train and run increasingly large models. The development of large campuses also places pressure on local power grids, water resources, and permitting processes. Investors are increasingly demanding data on cooling efficiency, renewable energy share, and flexible demand management capabilities.
Advanced power distribution systems and backup power also play a key role, with backup power required to maintain operations during grid failures. In the long term, the competitiveness of AI companies will depend equally on models, chips, networks, and the ability to secure reliable physical infrastructure.
