Sunrun and SPAN Partnered for Residential AI Compute
The companies are deploying liquid-cooled compute nodes in new homes to bypass traditional power transmission limits.
Updated on Sept. 23, 2026 in Data Centers

Live Poll
Would you allow a private company to host AI compute infrastructure in your home for compensation?
Sunrun and SPAN have announced an expansion of their partnership to integrate distributed AI compute infrastructure into new residential communities. The initiative utilizes solar-plus-storage systems to power SPAN's XFRA compute nodes, which feature liquid-cooled NVIDIA GPUs.
Why it matters
This partnership addresses the power generation and transmission constraints that currently throttle the deployment of large-scale AI data centers. By moving compute directly to the residential energy edge, the firms aim to solve the speed-to-power bottleneck limiting AI infrastructure.
The XFRA platform relies on a Secure Orchestration Layer to schedule AI workloads based on real-time energy availability and latency requirements. The infrastructure uses liquid-cooled NVIDIA GPUs to manage thermal loads within the residential hardware environment.
The players
Sunrun
A provider of residential solar energy systems and home battery storage solutions.
SPAN
A developer of home energy management systems and distributed compute infrastructure.
NVIDIA
A semiconductor company focused on designing graphics processing units and hardware for high-performance AI compute.
The details
The platform functions by combining Sunrun's home energy generation and battery storage systems with SPAN's orchestration software. The XFRA compute node, which integrates liquid-cooled NVIDIA graphics processing units — chips specialized in the high-speed parallel calculations required for AI — acts as a local data center. Homeowners participate via a pre-paid lease for a smart electrical panel and solar-plus-storage system, receiving ongoing compensation for hosting the compute hardware.
Timeline
September 23, 2026: Sunrun and SPAN announced the expansion of their partnership.
The Tech Race
This development moves beyond the industry-standard model of centralized hyperscale data centers that rely on massive utility grid access. It signals a shift in the race to secure AI compute capacity by leveraging the decentralized energy generation already present at the household level.
Customers in initial Texas deployments will receive smart panels and solar-storage systems via pre-paid leases. Participants will receive ongoing financial compensation for hosting the compute infrastructure within their homes.
The takeaway
This partnership tests whether residential solar arrays can act as reliable power plants for high-intensity AI workloads. Interested observers should monitor the first residential community deployments in Texas for performance benchmarks on compute uptime and grid reliability.
Further reading
For broader trends in infrastructure, explore the latest Data Centers analysis.
Live Poll
Would you allow a private company to host AI compute infrastructure in your home for compensation?









