Business
TeraWulf Partners with Fluidstack for Major AI Hosting Agreement

EASTON, Md., Aug. 14, 2025 (GLOBE NEWSWIRE) — TeraWulf Inc. has announced two significant 10-year high-performance computing (HPC) colocation agreements with Fluidstack, aimed at delivering more than 200 megawatts (MW) of critical IT load at its Lake Mariner data center campus in Western New York.
These agreements, projected to generate approximately $3.7 billion in revenue over the initial term, have potential total revenues of up to $8.7 billion if extension options are exercised. The deal includes two five-year extension options, providing TeraWulf with a robust revenue stream for the future.
To support the financing, Google will backstop $1.8 billion of Fluidstack’s lease obligations, obtaining warrants for about 41 million shares in TeraWulf, which represents an 8% equity stake. This strategic partnership indicates Google’s confidence in the project’s success and TeraWulf’s capabilities.
The deployment plan includes 40 MW set to go online in the first half of 2026, with the full capacity of over 200 MW expected to be operational by the end of that year. TeraWulf’s Lake Mariner facility is equipped for liquid-cooled AI workloads, ensuring efficient performance for demanding computing tasks.
“This is a defining moment for TeraWulf,” said Paul Prager, the company’s CEO. “We are proud to unite world-class capital and compute partners to deliver the next generation of AI infrastructure, powered by low-cost, predominantly zero-carbon energy,” he added.
Fluidstack’s co-founder, César Maklary, expressed pride in their partnership with TeraWulf, noting that it reflects their shared commitment to developing scalable infrastructure for AI. The involvement of notable partners like Google enhances TeraWulf’s standing in the market for AI infrastructure.
The Lake Mariner facility features advanced specifications including dual 345 kV transmission lines, closed-loop water cooling systems, and ultra-low-latency fiber connectivity, meeting the requirements for high-density AI clusters. With a projected cost of $8-$10 million per MW, TeraWulf aims to become a competitive player in the high-performance computing space.