Forget Iceland-style volcanic geothermal. The geothermal gradient is roughly 25–30 °C per kilometer everywhere on Earth — drill deep enough and any city sits on a furnace. What if every neighborhood drilled its own 5 km borehole pair and tapped the granite basement for district heating?

The thermal resource. At 5 km depth in typical continental crust, rock temperature reaches roughly 15 °C + 5 × 28 °C/km ≈ 155 °C. That's superheated water territory at modest pressure. The rock itself is the battery: granite has a heat capacity of about 2.2 MJ/(m³·K). Cool a cubic kilometer of 155 °C granite by 50 °C and you extract:

2.2e6 J/(m³·K) × 1e9 m³ × 50 K = 1.1e17 J ≈ 31 TWh

For comparison, a city of 100,000 homes burning 80 GJ/year each for heat needs 8 PJ ≈ 2.2 TWh annually. One cubic km of hot rock holds ~14 years of heating for that city. Drill a grid of borehole pairs spaced 500 m apart and you'd cover a neighborhood with decades of supply per "harvest cycle."

The catch is permeability. Granite has natural permeability around 10⁻¹⁸ m² — essentially impermeable. You can't pump water through it. Enhanced Geothermal Systems (EGS) solve this by hydraulically fracturing the rock to create a network of cracks, then circulating water between an injection and a production well. The Utah FORGE project (still running in 2026) and the Finnish ST1 Otaniemi well (5 km under Helsinki) have demonstrated the technique, but with a problem: induced seismicity. Basel, Switzerland abandoned its EGS project in 2009 after a magnitude 3.4 quake. Pohang, South Korea's 2017 magnitude 5.5 — the country's worst quake on record — was traced directly to EGS injection.

Drilling economics. A 5 km well in hard crystalline rock costs roughly $8–15M today. A neighborhood of 1,000 homes would need maybe two well pairs plus a heat exchanger plant — call it $50M of capex, or $50k per home. Spread over 30 years that's $1,700/year before financing, versus ~$1,500/year for natural gas heat in a cold climate. Marginal — until you price carbon.

The flow rate constraint. Each well pair can sustainably circulate roughly 50 L/s of water with a 50 K temperature drop. That's 50 kg/s × 4180 J/(kg·K) × 50 K ≈ 10.5 MW of thermal power per pair. A neighborhood peak heating load on a -10 °C night might hit 15 MW for 1,000 homes (15 kW peak per house), so you'd need two pairs running flat-out plus thermal storage for diurnal peaks. Hot water tanks the size of a swimming pool per block would do it.

The real engineering challenge isn't drilling — it's the heat exchanger. Deep geothermal brine is corrosive (silica, chlorides, dissolved CO₂) and scales the moment you drop temperature. Titanium plate exchangers and reinjection of fully-cooled brine are mandatory. Get the chemistry wrong and your $50M asset clogs in 18 months.

So: technically feasible, geologically universal, economically borderline, and seismically nervy. The cities that will do it first are the ones that already drill — Calgary, Houston, Stavanger — where the rigs and roughnecks are already on payroll.