Data centers are eating up land, driving up electricity costs and becoming lightning rods for public discontent with the power of big technology in society.
The Maine State Legislature recently passed a bill banning data centers in the state (but was unable to override the governor’s veto). As public opinion toward AI becomes increasingly negative, 14 states across the political spectrum from Oklahoma to New York are considering legislation to ban or suspend new data centers, according to the National Conference of State Legislatures.
Still, despite concerns from the public and politicians, funding to build new data centers is rapidly increasing. America’s largest technology companies are on pace to spend as much as $1 trillion a year on AI by 2027, according to recent Wall Street estimates. A recent McKinsey report predicts that worldwide data center spending will reach $7 trillion by 2030.
At the same time, the idea of locating data centers closer to consumers, even inside their homes, is gaining traction in the real estate industry. Major housing companies including house builders pluto groupis in initial testing Nvidia As CNBC’s Diana Orrick recently reported, California-based startup Span plans to install small fractional data center “nodes” on the exterior walls of new homes.
Whether that model is scalable and whether homeowners, HOAs, and regulators will approve of it is up for debate. Experts point out that home-based data centers have several advantages: Home-based grids reduce the need for new data center construction and improve energy efficiency.
“It’s technically possible and it’s already being considered,” said Balaji Tammabattula, chief operating officer of BaRupOn, a US-based energy and technology company currently building a data center campus in Liberty County, Texas. He said that just as home computers can provide processing power to distributed networks, homes can also host the computing hardware that powers large-scale data processing systems.
Advocacy groups and community members protest against data center laws outside the Texas State Capitol on Monday, February 23, 2026, in Austin.
Austin American Politician/Hearst Newspapers | Hearst Newspapers | Getty Images
The home-as-data center model would follow similar efforts to use latent home electricity for cryptocurrency mining, or to sell surplus rooftop solar power or EV credits.
“Feasibility depends on available power, internet connectivity, thermal management, and workload type. For batch processing and non-time-sensitive tasks, a home environment works surprisingly well,” Thammabatula said, but for dense AI training and real-time workloads, the constraints of a home environment are more difficult to overcome.
As thermal waste from data centers is receiving more attention as an issue in Europe, a real-world example is currently being deployed as a proof of concept. For example, a UK-based startup called Heata installs servers in people’s homes that handle cloud computing workloads and sends the heat generated directly to hot water cylinders in their homes, effectively providing homeowners with free hot water in exchange for hosting the hardware. British Gas supports testing this model.
On a larger scale, The operation has just begun Waste heat from Microsoft data centers in Finland will be used to power heat pumps that will heat the homes of approximately 250,000 local residents.
“These examples show that this concept works at both the household and community level,” Thammabatula said.
Home data centers come with a ledger of pros and cons. On the positive side, the residential model reduces land and infrastructure requirements, which are serious bottlenecks, and distributes computing closer to end users, creating a natural incentive for homeowners through energy savings, Thammabatula said. He added that home computing also has a strong sustainability perspective, as waste heat is recycled rather than being cooled at great expense.
But your questions for ChatGPT or Claude aren’t likely to be generated from a server in someone’s walk-in closet or basement anytime soon, and deep interactions with AI still require sprawling data centers. Currently, residential environments lack the power density, redundancy, physical security, and environmental controls needed for enterprise workloads. And if you can’t receive a WiFi or phone signal, you can’t power your data center.
“Connection quality varies from home to home, creating massive reliability issues. There are also regulatory and insurance issues when it comes to hosting commercial equipment in private homes,” Thammabatula said.
Currently, economics only works for certain workload types, such as batch processing, rendering, and research computation. “Anything that requires guaranteed uptime or low latency is not yet suited to this model,” he added.
Comparing home-based data centers and hyperscalers
Given their limitations, home data centers are much more likely to become a niche layer of future infrastructure than a replacement for hyperscale data centers. The home data center model typically involves a third party owning and operating the equipment, so the homeowner technically doesn’t have to manage anything.
“Hyperscale data centers will never replace homes, especially for large AI training clusters that require dense power, high-speed networking, specialized cooling, and tightly controlled environments,” said Gerald Ramdeen of Luxcore, a company developing next-generation optical networking and distributed cloud infrastructure. A more realistic opportunity, he said, is to turn homes into professionally managed edge computing nodes useful for AI inference, low-latency workloads, flexible/batch computing, cloud gaming, and certain heat recycling applications.
This approach will have an impact on our daily lives as we increasingly interact with and rely on AI in our daily lives.
“This can be used to sort through the 700 million photos that my teenage daughter has,” said Sean Farney, vice president of Americas data center strategy at JLL, a U.S.-based global professional services and commercial real estate company that manages 4.4 gigawatts of data center space worldwide from more than 340 data center sites.
Farney pointed out that while smartphones have more computing power than the first data center ever built, the idea of home data centers hasn’t caught on at scale yet, but it probably will. “It’s difficult to compete with hyperscalers because there are operating costs to maintain a hyper-distributed footprint. But it can be done, and the companies that do it right are looking at the right size valuations,” he said.
Home data centers still have some technical limitations before they can succeed on a commercial scale. First, Furney says data centers will quickly outstrip residential power supplies, so homes need a fairly reliable supply of electrical and mechanical resources. “A 20-kilowatt home generator won’t even get you an AI server cabinet,” he said.
But if technology can address these issues, can households overcome the scale effect of data centers? Fernie believes the answer is yes.
AI cybersecurity and physical security are challenges
Amy Simpson, director of product marketing at Huntress, a global cybersecurity company, said cybersecurity vulnerabilities are one reason to be skeptical about the widespread use of home-based data centers.
“The collection of home-based micro data centers creates a need for a more robust network security approach,” Simpson said. While home-based networks operating at scale may have the benefit of decentralization, with more sites comes greater redundancy in case one of your data centers goes down, a larger footprint also complicates security.
“Each site’s hardware and software must be secure and carefully monitored to avoid vulnerabilities,” Simpson said. The physical security of the site, on the other hand, “would be nearly impossible to guarantee,” she said. “There’s a reason why huge data centers run by companies like Amazon and Microsoft are surrounded by high fences and guarded 24/7.”
The Microsoft data center campus currently under construction is shown on September 18, 2025 in Mount Pleasant, Wisconsin.
Audrey Richardson | Reuters
“I can’t imagine a world where end users with data security and compliance obligations would be comfortable with the idea of their sensitive information being processed and managed by servers that might be sitting in someone’s garage,” Simpson said. Still, she knows of legitimate networks of micro data centers that use tamper-proof physical containers. If these could be installed inside the home, security concerns could be alleviated.
According to Arthur Ream, a lecturer in computer information systems at Bentley University, the model of using your home as a data center is plausible, is already happening, and is a sensible answer for inference, if not training, workloads.
“The interesting question is not whether home computing will work, but whether the security, reliability and regulatory narrative holds true at gigawatt scale, or whether the industry quietly understands that the cheapest place to avoid operational risk for AI is in someone else’s utility room,” Ream said.
Ream said Span is pioneering this model, with examples like its work with Nvidia and PulteGroup, where Span owns and installs liquid-cooled Nvidia RTX PRO 6000 Blackwell GPUs in homes and sells that compute to hyperscalers and AI cloud providers, allowing homeowners to get Span smart panels, battery backup, and discounted rates for power and internet. Homeowners pay a monthly fee of about $150 that covers electricity and internet. While SPAN sells compute to AI customers, installation is free.
“The economic argument should be taken seriously. A 100 MW data center costs about $15 million per megawatt and takes three to five years to build. They claim they can match that capacity by deploying XFRA nodes in 8,000 new homes. Even with aggressive cuts for marketing calculations, the difference in speed and power is real,” Ream said.
Other experts are less cautious, saying the concept won’t work.
“AI infrastructure is not cryptocurrency infrastructure. You don’t run data centers in your basement,” said Svyat Durianinov, chief strategy officer at San Francisco-based software and robotics company Bright Machines. Modern AI runs on “AI factories” of thousands of GPUs working together, and requires complex engineering, precision manufacturing, and tightly integrated supply chains, from building servers and racks to deployment. “Industrial-scale power and cooling will also be required. Computing will be moved closer to the edge, but in standardized, engineered systems rather than in crowd-sourced home data centers,” Duryaninov said.
Real estate experts are paying close attention to developments as data centers draw the ire of local communities, but they have their own concerns about how residential neighborhoods will react.
“HOAs will definitely be on board with this idea,” said Jeff Lichtenstein, president and founder of Echo Fine Properties in Palm Beach Gardens, Florida. “I can’t imagine our Facebook community pages,” Lichtenstein said. “A fight between a data company, a city, and a homeowners association would make a typical Republican-Democrat battle look like child’s play.”
