Overview
- Sunrun announced the pilot Wednesday and will deploy AI inference compute nodes inside participating homes that already have its solar panels and battery storage.
- The program will use Sunrun’s installed base of about 1.1 million residential systems and will pay homeowners for hosting the compute hardware.
- Compute nodes pair with home batteries so they can keep running during some outages and reduce the need for new data‑center siting and utility interconnection.
- Sunrun says it is in talks with enterprise offtakers to sell inference capacity while the company runs a multi‑month pilot and evaluates performance, homeowner experience, and rate‑structure effects.
- The move responds to fast growth in inference demand and complements Sunrun’s 16 GW flexible‑capacity framework with Tesla and Renew Home, but regulatory, grid and commercial economics still need validation before wider rollout.