Your solar panels and home battery are already working to lower your monthly utility bill. Now, Sunrun wants to turn that same hardware into a miniature data center. On July 8, 2026, the company launched a pilot program that places AI inference compute nodes directly into homes already equipped with their solar and storage systems. For the privilege of hosting this hardware, homeowners are looking at a potential payout of roughly $1,000 per year, according to estimates from Wells Fargo. With more than 1.1 million customer homes in its existing base, Sunrun has a massive potential footprint for this rollout.
The economic tension here is sharp: AI compute is incredibly lucrative, and the industry is desperate for space. Wells Fargo analyst Praneeth Satish notes that while traditional battery uses generate revenue well below $1 per kilowatt-hour, AI compute can command $4+ per kilowatt-hour. As Paul Dickson, Sunrun President and Chief Revenue Officer, put it: “We are now using our leadership position in distributed home energy and proven infrastructure to bring compute closer to the sources of energy and inference.”
This push arrives as McKinsey projects that inference demand will grow by about 35% annually, eventually surpassing training as the dominant AI workload by 2030. But while the industry is hungry for space, the public is increasingly wary of the infrastructure required to feed it. A March 2025 Gallup poll found that 70% of Americans oppose building an AI data center in their local area, with 48% expressing strong opposition. This community pushback is not just noise—it has already blocked or delayed $98 billion in projects, with at least 25 data centers canceled in 2025 alone. Even regulators are stepping in; as of July 2026, the New York Legislature has passed a moratorium on new hyperscale data centers with 20MW+ peak demand.
Homeowners considering this trade-off face immediate security realities. Security expert Anthony Saquella warned in May 2024 that residential AI compute nodes could create risks similar to the Mirai botnet. Because home networks typically lack the rigorous segmentation found in enterprise environments, these nodes could introduce lateral-movement risks, potentially exposing your personal devices to the wider internet. There is also the question of liability: if a node hosted in your home is linked to illegal activity or becomes the subject of a cyber investigation, the homeowner could find themselves in a difficult legal position.
Sunrun is not the only player testing this model. A partnership between Nvidia, PulteGroup, and Span is planning to install 100 residential compute nodes in Nevada and Arizona later this year. That program offers a much higher incentive, with homeowners paid approximately $22,000 per year to host Nvidia Blackwell GPUs. Compared to that, the $1,000 estimate for the Sunrun pilot feels like a different tier of commitment, though it is important to note that Sunrun has not officially confirmed any compensation terms.
In fact, the Sunrun pilot remains remarkably opaque. As of mid-July 2026, there are no public technical specifications regarding GPU models, thermal management, bandwidth, or latency benchmarks. We don’t know who the enterprise compute customers are, and the pilot is still in its early stages. While Sunrun’s stock rose about 7.3% in premarket trading following the announcement, the actual experience for the homeowner remains a black box.
People are worried about their own bottom lines, too; Consumer Reports found in November 2025 that 78% of U.S. adults were concerned that new data centers would drive up their energy bills. The decision to host an AI node is a gamble on the future of distributed infrastructure. You are essentially deciding whether the potential for a modest annual payment outweighs the risks of network exposure and the unknown technical demands of running a server in your home. As the industry races to find space for the next wave of AI, your home is being positioned as the next layer of the compute stack. Whether that is a smart investment or a security headache is something we will only learn as the pilot progresses.