raw · papers · ingested 2026-09-19
2025 Geothermal Drilling Cost Curves Update (NREL / Stanford Geothermal Workshop 2025)
Source: https://docs.nrel.gov/docs/fy25osti/92793.pdf
Provenance: NREL conference preprint (Stanford GW 2025), DOE-GTO funded, using Fervo-supplied well data plus Utah FORGE. Independent government-lab quantification of the EGS drilling learning curve that feeds the GETEM / SAM / ATB cost models. Retrieved via research agent; full text extracted with pdftotext.
Why drilling matters
“Drilling activities account for 30% to 57% of the cost to develop and install a geothermal plant.” The share is higher for next-generation / EGS projects.
Per-well record (El-Sadi et al. 2024 data)
| Well | Days | TD (ft) | $/ft | Note |
|---|---|---|---|---|
| Project Red well 1 | 71 | 11,220 | 994.62 | 158 ft/day |
| Project Red well 2 | 58 | 11,211 | 1,106.74 | |
| Cape well 3 | 33.75 | ~13,300–13,950 | 633 | |
| Cape well 4 | 25.91 | 472 | ||
| Cape well 5 | 24.26 | 396 | ||
| Cape well 6 | 20.1 | 346 | ||
| Cape well 7 | 23.71 | 394 | ||
| Cape well 8 | 22.27 | 357 |
“the Project Red well costs were close to three times the average cost of the Cape Station wells.” Inter-project learning rate ~35%.
Cape Station drilling facts
- 1 vertical + 14 horizontals from just two pads (Frisco, Bearskin), Jun 2023–Sep 2024; TD 13,272–15,347 ft (4,045–4,678 m); laterals ~4,700 ft; max temperature 444 °F.
- Triplet test: producing 382.9 °F (195 °C), peak >12 MWe, sustained 8–10 MWe.
- Average ROP 70 ft/hr with 34 hr average bit life (PDC bits). Compare Utah FORGE 16B(78)-32: 102 ft/hr but 6.5 hr bit life (16 PDC bits, 10,947 ft, 75 days); Geysers GDC-36 65 ft/hr. Single-run hard-rock ROP up to 173 ft/hr at FORGE.
Model assumption changes vs GeoVision (2017)
- ROP 25 → 75 ft/hr; casing-running speed 300 → 800 ft/hr (casing running tools); contingency 15% → 10% (Cape cumulative cost was 84% of AFE); PDC bit cost multiplier 2×.
- Result: drilling costs 12–24% below the GeoVision baseline for vertical wells, 18–26% for horizontal wells; curves now sit between GeoVision “Baseline” and “Intermediate 1.”
- Drivers: PDC bits, physics-based limiter redesign (mechanical specific energy), multiwell pad drilling, oil-and-gas technology transfer.
- Caution: a 2022 attempt to lower the curves failed because Utah FORGE costs did not match its ROP gains — speed ≠ cost.
Key quotes
- “For vertical wells, the magnitude of cost reduction ranges from 12% to 24%, while for horizontal wells, the estimated cost reduction is between 18% and 26%.”
Companion: Fervo drilling blog (12 Feb 2024)
https://fervoenergy.com/fervo-energy-drilling-results-show-rapid-advancement-of-geothermal-performance/ — fastest Cape well 21 days (70% faster than the first Red well); cost per well $9.4M → $4.8M across the first four Cape horizontals; Cape wells 2,100+ ft deeper than Red; ~70 ft/hr ROP; planned 18% learning curve, achieved ~35%.
Relevance to the cooling wiki
Fixes the cost side of “geothermal as drive heat”: the well is the dominant cost, and the well cost is now a moving target ($9.4M → $4.8M in six months at one site). Any levelized cost of geothermal cold inherits this curve.