Sunrun, the largest residential solar and battery storage company in the US, just announced something that sounds like it was cooked up in a crypto whitepaper circa 2021: turning people’s homes into a distributed network of AI data centers. The difference is that Sunrun has 1.1 million home systems already installed and a NASDAQ ticker (RUN) to go with the vision.
The company is launching a pilot program that will place AI compute nodes inside homes already equipped with Sunrun solar panels and battery storage. Homeowners get compensated for hosting the hardware. Sunrun sells the resulting compute power to enterprise clients, particularly AI companies hungry for inference capacity.
How distributed AI compute actually works here
Sunrun’s pilot will install compute units in participating customers’ homes, tapping into the solar and battery infrastructure already in place to power them. These nodes handle AI inference workloads, the kind of computing that runs trained AI models rather than training new ones.
McKinsey projects AI inference workloads will grow by roughly 35% annually, potentially surpassing training workloads as the dominant use case by 2030. That’s the market Sunrun is chasing.
