According to the announcement shared with Bitcoin.com News, the sale, which includes $85 million in cash and mining equipment valued at $7 million, represents a 3.4x return on Terawulf’s initial investment. The transaction is seen as part of a strategic shift to focus on more efficient and scalable AI and HPC operations. Terawulf aims to reinvest the capital to construct a new 20-megawatt colocation building (CB-1), with operations expected to begin in early 2025.
“This transaction further aligns Terawulf’s focus and investments with where we have the most operational efficiency, the greatest growth potential, and the best opportunity to drive incremental value for shareholders,” Paul Prager, the chief executive officer of Terawulf remarked.
In addition to the AI-focused expansion, Terawulf insisted it remains committed to maintaining its bitcoin mining capabilities. The company plans to upgrade its fleet’s efficiency at the Lake Mariner facility, targeting a capacity of 13 exahash per second (EH/s) by the first quarter of 2025. Alongside this, Terawulf disclosed it is working to complete another colocation building, CB-2, which will have 50 megawatts (MW) of capacity and is expected to be operational by mid-2025.
Many bitcoin miners had already been shifting toward HPC and AI services even well before the fourth Bitcoin halving reduced their block rewards. Since the halving, mining revenue has significantly declined, prompting miners to accelerate their transition to AI services while also mining other proof-of-work assets, such as kaspa and fractal bitcoin. Given their expertise in running large-scale data centers, providing HPC and AI hosting services has become their natural progression.
Terawulf’s decision to monetize its Nautilus stake comes ahead of the expiration of a low-cost power agreement with Talen Energy in 2027. By divesting its interest now, Terawulf further detailed that it intends to capitalize on better cost efficiencies at its New York facility while ensuring that it can meet the growing demand for AI and HPC services using predominantly zero-carbon energy sources.
