Sunrun is expanding beyond solar and home batteries with a pilot for distributed AI compute, according to The Verge. Instead of building a conventional data center, the company plans to place compute nodes in participating customers' homes and compensate them.
The idea connects two major infrastructure pressures: AI's appetite for compute and the energy constraints around large centralized data centers. A distributed model could use residential energy assets differently, though reliability, maintenance, networking, and customer acceptance remain open questions.
If the pilot works, it could broaden the conversation about where AI infrastructure can live and who participates economically in hosting it.