Sources: Meta's “Project Walleye” Ohio data center seeks $3B in loans in a first-of-its-kind deal where lenders will fund both the building and the power assets
‘Project Walleye’ lenders would be first to fund both construction and power — A data centre campus backed …
Context & Ripple Effects
Meta has already been shifting large AI-campus funding beyond a conventional on-balance-sheet build: its Hyperion joint venture with Blue Owl paired outside capital with Meta retaining an ownership stake, while a separate Indiana campus was planned as another large-scale buildout. Project Walleye extends that financing experimentation to the power infrastructure needed to make a campus usable.
The deal also arrives as investors pursue control of data-center sites with secured electricity, exemplified by Silver Lake's powered-land development effort. Combining the building and power assets in one loan package makes grid access part of the financeable project rather than a separate development dependency.
First-order effects
- Meta can seek debt for both campus construction and associated power assets in a single Ohio financing, potentially reducing the need to coordinate separate capital structures for each layer.
- Prospective lenders must underwrite not just a data-center building but the power assets and their delivery risk, broadening the collateral and operational exposure attached to the loan.
Second-order effects
- If funded, the structure gives other hyperscale-campus sponsors and lenders a concrete precedent for packaging power alongside real estate, particularly where electricity availability constrains development.
- Capital providers and developers with control of powered land may gain leverage: the value of such sites becomes more directly tied to whether power assets can be financed and delivered with the facility.
Third-order effects
- AI-infrastructure finance is likely to become more asset-specific, with power availability and delivery risk increasingly embedded in loan terms, valuation, and project governance rather than treated as a background utility cost.
- If this approach is adopted more broadly, data-center expansion could depend as much on lenders' appetite for integrated power-and-compute risk as on operators' capital spending plans; that could concentrate development around financeable power arrangements.
The trend: This is one data point in the financialization of AI compute, in which hyperscalers use specialized outside capital to fund the buildings, energy assets, and execution risks behind new capacity.