What if a neighborhood shared one heating and cooling system, the way it already shares water mains and power lines?
Since June 2024, a few streets in Framingham, Massachusetts, have been finding out. Eversource, the state's largest utility, turned on what it calls the country's first utility-owned networked geothermal system: an underground loop connecting homes, apartments, and businesses so they can all draw heat from the ground. It runs through a New England winter, which is the part that makes people pay attention.
How a shared loop works
A few hundred feet down, the ground holds at roughly 55 degrees Fahrenheit all year. That's not warm enough to heat a house directly. But a heat pump can take that steady warmth, concentrate it, and push it into a building at a comfortable temperature. In summer it runs the other way, pulling heat out of the building and sending it back into the ground. One system handles both heating and cooling.
A single home can do this alone with its own boreholes, but drilling is expensive. One homeowner interviewed by WBUR was told the borehole for her backyard alone could cost more than $40,000. The network idea is to drill once for many buildings and run shared pipes under the street, so each building only needs a heat pump and a connection.
Framingham's loop serves 36 buildings on its planned route, 24 of them homes, according to Eversource. Counting each apartment and business, that's about 140 residential and commercial customers, by GCN's count. The pilot runs through Eversource's gas business. One thing it is testing, beyond the engineering, is whether a company that owns pipes under the street could deliver heat instead of fuel.
Who pays what
This is the part I find most interesting.
Under the pilot, Eversource paid to build the network and to install the heating and cooling equipment inside participating buildings. Participants didn't get the usual installation bill. Instead, residential customers pay a fixed charge of about $10 a month to connect, and income-qualified households pay about $7, according to Eversource figures reported by Utility Dive. Each customer also pays for the electricity that runs their own heat pump.
The system cost roughly $15 million to build, by Utility Dive's account. Other reports put it at about $14.7 million. Divide that by around 140 customers and you get more than $100,000 per customer. That's my own rough arithmetic, not a figure Eversource has published, and a first-of-its-kind pilot carries design and learning costs a repeat project may not. Still, it shows how far the $10 monthly fee is from the cost of what's underground.
Someone pays the difference. GCN reported this month that regulators approved recovering the rest of Framingham's construction cost from Eversource's broader customer base. The cost didn't go away. It was spread across many more customers, most of whom aren't connected to the loop.
In my LinkedIn post I credited this finding to a 2025 analysis by Maryland's Office of People's Counsel. I couldn't locate that analysis to confirm it, so I've attributed the point to the reporting I could verify.
The price after the pilot
The $10 charge was a pilot price, and the pilot was set to run for 24 months, through two heating and cooling seasons.
In September 2025, Eversource asked Massachusetts regulators to approve a permanent framework for networked geothermal service. As Canary Media reported in May 2026, the proposed rate would charge a home with a three-ton heat pump a fixed $10 a month plus $14.95 per ton of capacity, for a total of $54.85 a month. As far as I could find, it was still awaiting approval in October 2026. Two things are worth noting. That rate is a proposal, not a final price. And it covers heat drawn from the ground with no fuel cost, so you can't compare it directly to a gas bill without the electricity cost and a household's actual usage. I don't have that full comparison, and I'd be careful with anyone who claims it in one number.
The counter-evidence
Not every attempt has gone ahead. In December 2024, National Grid canceled its planned geothermal network in Lowell, Massachusetts. A company spokeswoman told WBUR that "the cost per customer for this project was unfortunately much too high." She blamed inflation, supply costs, and difficult geology.
Maryland has run into the same tension. Washington Gas proposed a network for the Northwest Park apartments in Montgomery County with a flat $12.50 monthly charge for residents. GCN reports that the utility's own estimate had customers who weren't participating paying $5 to $8 more a month. The commission reviewing the plan pushed back on the numbers.
On the other side, Framingham's technical performance was steady as of early 2025, according to an Eversource spokesman quoted by WBUR. The Department of Energy awarded $7.8 million for a second phase, which GCN reports is partly designed to test whether connecting a new network to an existing one can bring costs down.
Pilot or prototype?
Here's what I keep thinking about. Is Framingham a financing model that could scale? Or is it mainly a test of whether the technology works, with the assumption that a sustainable way to pay for it comes later?
My read is that it's mostly the second, for now, and that's not a criticism. Plenty of shared infrastructure started out expensive per user and got cheaper as networks grew. Whether geothermal networks follow that path is exactly what the second phase and the permanent rate case are meant to show. Lowell is a reminder that sometimes the answer comes back no.
There's an idea here I've touched on before. In The Discount You Can't Earn Alone I wrote about fire mitigation that only pays off when a whole neighborhood does it together. Framingham is a version of the same idea for heating. Some upgrades are cheaper, or only possible, when a block does them together, and the hard question is always who fronts the money and how it's shared.
Today, replacing a furnace is usually a private decision a homeowner pays for. In Framingham, for about 140 customers, it became a utility service with a monthly fee and a regulator overseeing it. Whether that arrangement can grow without the rest of the rate base carrying it is the question the pilot was built to answer, and it hasn't answered it yet.
Sources
- Eversource, Geothermal Pilot in Framingham. Commissioned June 4, 2024; planned route of 36 buildings, 24 residential; Eversource pays for the network and all in-building equipment; participants pay a fixed monthly access charge plus their own electricity; 24-month pilot.
- Utility Dive, report on the Framingham system going live, June 2024. Roughly $15 million; about $10 a month residential and $7 income-qualified; ground temperature around 55°F.
- WBUR, "A project in Mass. tests a future for gas utilities without fossil fuels," March 8, 2024. Constant 55°F a few hundred feet down; backyard borehole cost "upwards of $40,000."
- GCN, "A Maryland pilot quoted homeowners a flat $12.50 a month for shared geothermal heat…," October 16, 2026. Framingham: about 140 customers; regulators approved recovering remaining construction cost from the broader customer base; $7.8 million DOE grant for a second phase. Washington Gas Northwest Park: $12.50 resident charge; utility estimate of $5–$8 a month for nonparticipating customers.
- Canary Media, "As geothermal networks grow, so does the call for a new utility model," May 28, 2026. Eversource's proposed post-pilot rate: $10 fixed plus $14.95 per ton; $54.85 a month for a three-ton system; awaiting regulatory approval. Framework petition docketed as Massachusetts D.P.U. 25-86.
- WBUR, "Gas utility cancels networked geothermal pilot in Lowell," February 6, 2025. National Grid canceled in December 2024; cost per customer "much too high"; Framingham technical performance steady.
- ThinkGeoEnergy, "Networked geothermal heating pilot to go online in Framingham, Massachusetts." Pilot expected to cost about $14.7 million.
Figures current as of October 2026.
