Sources: a group of banks is in talks to lend $15B to Nexus to build a Texas data center; Anthropic will lease it and Google has provided financial guarantees
Google's guarantees of power and lease obligations would help developer secure financing for the 1.6-gigawatt Texas project
Context & Ripple Effects
This is a financing step in a reported arrangement first outlined when Google was said to be nearing support for Nexus's Texas campus, with Anthropic as tenant. It follows Anthropic's reported move into more than a dozen initial direct data-center lease agreements, shifting its infrastructure commitments beyond a purely cloud-provider relationship.
The proposed structure also fits the reported scale of Anthropic's planned Google Cloud and chip spending: a large multiyear Google commitment gives the partnership a clearer commercial backdrop for third-party infrastructure financing.
First-order effects
- Nexus could gain access to debt for the Texas build if the bank talks result in a loan; Google’s reported guarantees are designed to make the project’s power and lease revenue more credible to lenders.
- Anthropic would secure a dedicated leased facility rather than merely consume general cloud capacity, while Google takes on contingent obligations tied to the project’s power and tenancy.
Second-order effects
- The arrangement makes credit support from a strategic cloud partner a more important tool for data-center developers seeking to fund AI-specific capacity, especially where a tenant alone may not satisfy lenders’ underwriting requirements.
- Other AI developers and infrastructure providers may face pressure to pair long-term capacity commitments with stronger guarantees or other risk-sharing terms when competing for large-scale construction financing.
Third-order effects
- If replicated, AI infrastructure could increasingly be financed as contracted capacity: developers build, model companies lease, and cloud platforms absorb selected credit or operating risks to bridge the two.
- That structure can accelerate buildouts, but it also concentrates execution exposure among tenants, guarantors, power arrangements, and lenders; the durability of the model will depend on whether those commitments hold through construction and operation.
The trend: AI compute is becoming a project-finance market in which leases, power obligations, and cloud-provider guarantees are packaged to make large facilities financeable.