cd /news/ai-infrastructure/nvidia-partners-with-homebuilders-to… · home topics ai-infrastructure article
[ARTICLE · art-17715] src=cryptobriefing.com pub= topic=ai-infrastructure verified=true sentiment=↑ positive

Nvidia partners with homebuilders to put AI data centers in residential backyards

Nvidia has partnered with smart electrical panel maker SPAN and homebuilder PulteGroup to deploy compact AI compute nodes, called XFRA, on residential properties. The pilot program targets around 100 nodes for 2026, tapping into underused household electrical capacity to ease AI's growing power crunch. Homeowners pay roughly $150 per month for electricity and internet, receiving infrastructure upgrades such as smart panels and battery backups in return.

read2 min publishedMay 29, 2026

A new pilot program deploys GPU-packed compute nodes on residential properties, tapping into underused household electrical capacity to ease AI's growing power crunch.

Nvidia has partnered with smart electrical panel maker SPAN and homebuilder PulteGroup to deploy compact AI compute nodes, called XFRA, on residential properties. The pilot program is targeting around 100 advanced nodes for launch in 2026, with a long-term vision of scaling past 1 gigawatt of distributed compute capacity.

What’s actually happening here #

Homeowners who participate typically pay a fixed fee of roughly $150 per month covering electricity and internet service for the node. In some cases, they may receive discounts or free services instead. The value proposition for the homeowner comes in the form of infrastructure upgrades: smart electrical panels, battery backup systems, and general enhancements to the property.

The XFRA nodes themselves are serious pieces of hardware. Each unit packs 16 Nvidia RTX PRO 6000 Blackwell GPUs alongside 4 AMD EPYC CPUs, with approximately 3 to 4 terabytes of RAM per node.

Marc Spieler, Nvidia’s senior managing director of global energy industry, has emphasized the potential of tapping into residential power capacity that often goes unused during large portions of the day. SPAN’s technology identifies and optimizes that slack, routing it toward compute workloads without disrupting the homeowner’s normal power usage. Residential electrical capacity is typically underused by as much as 58%.

The math behind the pilot #

The initial deployment targets southwestern US regions, aiming for approximately 1.2 megawatts of total capacity from those first 100-odd nodes. SPAN’s projections suggest that distributing nodes across roughly 8,000 homes could deliver the equivalent capacity of a 100 megawatt centralized data center, and they claim it could be done in approximately six months. The cost of deploying this distributed system runs about $3 million per megawatt.

Current testing is being conducted in PulteGroup-constructed homes. SPAN published a whitepaper on the concept in April 2026, followed by public announcements in May 2026. Traditional large-scale data center grid interconnections can take 4 to 7 years for large-scale projects.

The program has no connection to cryptocurrency or blockchain technology. The project remains in its prototype and pilot phase as of mid-May 2026, so real-world adoption data simply does not exist yet.

Disclosure: This article was edited by Editorial Team. For more information on how we create and review content, see our

Editorial Policy.

── more in #ai-infrastructure 4 stories · sorted by recency
sponsored brought to you by zahid.host 4,200+ EU-deployed projects
reading about agents? ship yours in a single git push.

Run your AI side-project on zahid.host

EU-based hosting, git-push deploys, automatic HTTPS, no cold starts. Free tier with a custom domain — perfect for shipping the agent you just read about.

$git push zahid main
Live at https://your-agent.zahid.host
Get free account → Pricing
from €0/mo · no card required
LIVE [news/nvidia-partners-with…] indexed:0 read:2min 2026-05-29 ·