- SPAN plans to install AI-powered GPU boxes outside ordinary suburban homes
- Homeowners offered subsidized electricity for hosting remote computing infrastructure equipment
- Each neighbourhood node contains sixteen expensive Nvidia GPUs inside compact enclosure
A San Francisco start‑up called SPAN has proposed placing small data center nodes outside suburban houses.
The company says it aims to install thousands of liquid‑cooled boxes called XFRA nodes, each containing powerful Nvidia GPUs.
Homeowners would receive subsidized or even free electricity and internet access in exchange for hosting this equipment on their property.
A quiet box with sixteen GPUs
Each XFRA node attaches to the exterior wall of a house like an additional utility box.
The unit holds sixteen Nvidia RTX Pro GPUs and runs with minimal noise, according to the company's announcements.
SPAN claims it can install eight thousand such nodes for five times less money than building a conventional data center with the same computing power.
“Data centers are loud, ugly, and often drive up local electricity bills,” said Chris Lander, vice president of XFRA at SPAN.
“This is quiet, discreet, and makes energy more affordable for the host and community.”
The system taps into excess electrical capacity that already exists in most modern American homes.
“Virtually all homes with 200‑amp utility services have 80 amps available at all times, so we set that as the maximum power consumption for a single XFRA node,” Lander explained.
“This home backup is provided to the host at no cost to them, contributing to greater energy resilience in addition to affordability,” he added.
Benefits for utility companies and communities
SPAN argues distributed nodes help grid operators avoid costly infrastructure upgrades, and that increasing electricity sales over existing grid infrastructure makes power more affordable for everyone.
The approach focuses on AI inference tasks rather than model training, which requires thousands of GPUs working together.
However, not everyone shares SPAN’s optimism. Ari Peskoe, a director at Harvard Law School, cautioned that utility companies may need to adapt their local grid management for neighbourhoods with many such nodes.
“If there’s a block that has several homes with these devices, maxing out compute and energy would force a lot of power to that local area,” Peskoe said.
However there are security concerns over the project, as thieves may also target these boxes, since each GPU sells for around $10,000.
The company plans a 100‑home pilot deployment in 2026, followed by rapid scaling to 80,000 units across the United States by 2027.
Whether suburban homeowners will accept this arrangement, which many may not understand, remains uncertain.
Meanwhile, the willingness of utility regulators and local zoning boards to approve such a decentralized computing experiment remains to be seen.
The pitch sounds appealing on paper, yet the real test will come when actual residents discover what it means to live next to a box of expensive electronics that strangers control remotely.
Follow TechRadar on Google News and add us as a preferred source to get our expert news, reviews, and opinion in your feeds.