Data-center developer TeraWulf has secured a major vote of confidence from one of the biggest names in tech. Alphabet's Google will provide a credit backstop for a new campus in Hawesville, Kentucky, that is being built around a 20-year lease with AI firm Anthropic. The lease is expected to generate roughly $19 billion in contracted revenue over its life.
The arrangement is a significant step for TeraWulf, which has been pivoting from bitcoin mining toward AI infrastructure. The Kentucky site is planned to reach about 401 megawatts of “critical IT load” — the power available to run servers — with initial capacity targeted for the second half of 2027.
Why a credit backstop matters
Financing large data-center projects is rarely straightforward. Even with a long-term lease in hand, lenders worry about whether the developer can raise enough capital and finish construction on time. That's where Google's involvement changes the picture.
By offering “investment-grade” credit support, Google is effectively telling banks they can underwrite some of the risk against Google's balance sheet rather than relying solely on TeraWulf's. This typically translates into cheaper borrowing costs, longer repayment timelines, and less pressure on the developer to issue new shares or other expensive equity to fund the build.
For a company like TeraWulf, which is smaller and less established than the tech giants it serves, that kind of backing can be transformative. It makes the project look more like a bankable deal — the sort usually reserved for larger, better-rated operators.
TeraWulf also expanded its agreement with Fluor Enterprises, an engineering and construction firm. That move signals the project is moving from announcements toward actual execution, a key checkpoint for investors who have seen many AI data-center plans stall at the drawing board.
What it means for investors
For markets, the Google backstop changes the financing math on a $19 billion lease. Capital costs are often the deciding factor in AI data-center projects: power hookups, land, and server-ready buildings require enormous upfront spending. A backstop like this can narrow the “credit spread” — the extra interest a borrower pays over safer debt — and open the door to longer-term project loans.
The flip side is competitive pressure. Developers and “neocloud” providers without a Big Tech guarantee may face structurally higher funding costs, even if they can line up long-dated customer contracts. That could widen the gap between the haves and have-nots in the AI infrastructure boom.
This is part of a broader trend where hyperscalers are increasingly willing to put their balance sheets behind the infrastructure they depend on. Google has been shielding its AI and cloud teams even as it trims elsewhere, and the company's appetite for data-center capacity shows no sign of slowing.
For everyday investors, the key takeaway is that AI's physical footprint — the buildings, power, and cooling that make models like Anthropic's possible — is becoming a bigger part of the investment story. Companies that can secure strong backing, like TeraWulf, may be better positioned than those that can't. But it's also worth remembering that these projects carry execution risk: construction delays, cost overruns, and shifting demand can all derail even the best-financed plans.
Investors will be watching for updates on the Kentucky build, including whether TeraWulf hits its 2027 target and how the financing terms shape up. The valuation discipline around Anthropic and the broader policy environment for data centers will also matter. And with AI data centers driving demand across the economy, deals like this one are likely to keep coming.
For now, the Google backstop is a clear signal that Big Tech is willing to put its money where its AI ambitions are. That's good news for TeraWulf, and a reminder that in the AI race, the balance sheet is just as important as the algorithm.
