This article first appeared on GuruFocus.

Google (GOOGL, Financials) is broadening its push into AI hardware by adopting financing tactics similar to those Nvidia has used to dominate the market for advanced compute. As demand for AI capacity continues to rise, Google has begun using its credit strength to help customers fund data-center projects designed for its Tensor Processing Unit chips.

• Warning! GuruFocus has detected 5 Warning Signs with AMD.

• Is GOOGL fairly valued? Test your thesis with our free DCF calculator.

The most recent example is a deal with TeraWulf and FluidStack, where Google provided a lease guarantee of up to $1.8 billion. The backstop allowed the partners to secure long-term, lower-cost financing for hundreds of megawatts of new AI infrastructure that lenders would not otherwise support. In exchange, Google obtained warrants equal to roughly 8% of TeraWulf and secured dedicated TPU-ready capacity.

The structure mirrors Nvidia’s own vendor-financing model, under which it has helped customers lease large GPU clusters and invested in AI firms that rely on its hardware. Both companies are increasingly pairing technology and financial engineering to lock in high-growth AI workloads.