American mining company TeraWulf has received credit support from Google as part of a project to build data centers for Anthropic, the developer of the Claude model. The credit support agreements will allow TeraWulf to raise debt financing against Anthropic's long-term lease obligations totaling about $19 billion.
The key asset in the deal is the Justified campus in Kentucky with a capacity of about 401 MW, which will serve Anthropic's computing needs. Capacity delivery is scheduled to begin in late 2027 and be completed in the second quarter of 2028.
Why this matters for the market
The deal demonstrates a sustained trend: mining companies with energy and infrastructure assets are increasingly pivoting to serve demand from AI developers. For TeraWulf, this means a transformation of its business model — from bitcoin mining to long-term contracts for the supply of computing capacity, which reduces revenue volatility and increases the predictability of cash flows.
Google's participation as a credit guarantor underscores the strategic importance of computing resources for leaders of the AI industry. Giants are willing to take on financial risks to secure access to large-scale capacity for years to come. The $19 billion in obligations testifies to the scale of the race for infrastructure, where energy assets are becoming no less valuable than the algorithms themselves.
For the crypto industry, this is a signal of further convergence between the mining and AI sectors. Companies able to offer ready-made sites with power supply gain access to capital that is incomparable to the traditional economics of mining.
Cryptalist analysis: The deal cements a structural shift in how mining companies are valued — the market is beginning to perceive them not as cyclical bitcoin miners, but as infrastructure operators with a long-term contract base. Google's participation reduces the project's credit risks and may set a benchmark for similar agreements in the industry. In the medium term, this intensifies competition for energy resources and supports a revaluation of the sector toward AI infrastructure.