TeraWulf’s 20-year lease with Anthropic sets a new reference point for revenue-backed AI infrastructure, combining investment-grade credit with a 401 MW campus footprint in Kentucky. The arrangement effectively underwrites a multi-phase build plan while aligning capacity timing with Anthropic’s model roadmap and training cadence. Pairing this with the sale of TeraWulf’s 50.1% stake in the Abernathy JV to Fluidstack recycles capital into wholly owned assets, simplifying governance and consolidating economics. The through line: longer-duration contracted revenue, fewer joint-venture constraints, and clearer control of customer relationships—critical when power allocation, interconnection timing, and cooling design can make or break AI campus ROI.
The Justified Data campus targets initial service in the second half of 2027, ramping to 401 MW by early 2028. That cadence dovetails with expected node shifts, liquid-cooling adoption, and rising rack densities that push 80–150 kW per rack in training clusters. The implied engineering challenge is not only power acquisition but also distribution, substation reliability, heat rejection, and water stewardship. Locational advantages—including transmission access and siting for modular capacity additions—let operators commission in phases and match energy procurement to compute waves. Anthropic’s long-term commitment reduces stranded-capex risk across switchgear, transformers, and cooling assets while improving the amortization profile for heavy electrical infrastructure.
Financially, the ~$19B contracted revenue base can support project-level debt and lower the weighted average cost of capital versus merchant capacity. This enables TeraWulf to prioritize high-certainty expansions and pre-order long-lead electrical equipment with improved pricing and delivery slots. Exiting the Abernathy JV at a premium monetizes an estimated ~$450M investment and reduces financial statement complexity, shifting from equity-method exposure to direct ownership returns. The net effect is a more bankable cash flow stack, potentially enabling structured financings such as lease-backed notes, securitized receivables, or hybrid instruments that reward duration and credit quality from a top-tier AI tenant.
For AI builders and enterprise buyers, the signal is clear: capacity certainty now hinges on pairing megawatt-scale leases with credible energy roadmaps and liquid cooling-readiness. Procurement teams should evaluate campus partners by their track record in grid interconnections, substation execution, and thermal envelopes aligned to next-gen accelerators. For investors, the trade is moving from speculative megawatts to credit-secured, phase-gated campuses that de-risk expansion. The Kentucky campus timeline also underscores the industry’s gating factor—power and infrastructure lead times—making early commitments, demand forecasting, and flexible module design essential to deliver contiguous training windows without productivity-killing migrations.


