Skip to content
gentic.news — AI News Intelligence Platform
Connecting to the Living Graph…

Listen to today's AI briefing

Daily podcast — 5 min, AI-narrated summary of top stories

UPTIME RELIABILITY COMPARISONBEFORE25%SolarAFTER90%Geothermal +260.0% deltagentic.news
Auto-generated diagram from article data — Uptime reliability comparison

DOE Backs Geothermal for AI Data Centers: 8 Pilot Sites Funded

DOE funds 8 geothermal pilot projects to power AI data centers, targeting 5 GW by 2030. Geothermal offers 90%+ uptime, bypassing grid delays.

·3d ago·3 min read··3 views·AI-Generated·Report error
Share:
Source: news.google.comvia gn_dc_powerSingle Source
How is the DOE using geothermal energy to power AI data centers?

DOE funded 8 geothermal pilot projects to power AI data centers, targeting 5 GW of firm clean power by 2030. Geothermal provides 24/7 baseload with 90%+ uptime vs solar's ~25% capacity factor.

TL;DR

DOE funded 8 geothermal pilot sites for AI data centers · Geothermal offers 24/7 clean power with 90%+ uptime · AI data center load expected to hit 35 GW by 2030

DOE funded 8 geothermal pilot projects to power AI data centers, targeting 5 GW of firm clean power by 2030. Geothermal offers 24/7 baseload with 90%+ uptime vs solar's ~25% capacity factor.

Key facts

  • DOE funded 8 geothermal pilot projects for AI data centers
  • Target: 5 GW firm clean power by 2030
  • AI data center load projected at 35 GW by 2030
  • Geothermal provides 90%+ uptime vs solar's ~25%
  • Total investment: $165M DOE + $215M industry match = $380M

The Department of Energy announced $165 million in funding for 8 geothermal pilot projects specifically designed to co-locate with or directly power AI data centers. The initiative targets 5 GW of firm clean energy capacity by 2030, a fraction of the projected 35 GW AI data center load, per DOE estimates.

Why geothermal now
AI data centers consume 50-100 MW per facility, with some planned campuses exceeding 1 GW. Intermittent renewables (solar ~25% capacity factor, wind ~35%) struggle to match the 24/7 reliability that GPUs demand. Geothermal provides 90%+ uptime with zero carbon emissions, making it the only renewable that can replace gas peaker plants for AI base load.

The 8 pilot sites span California, Nevada, Utah, Oregon, Idaho, New Mexico, Texas, and Alaska. Each project pairs a geothermal developer with a data center operator (names not disclosed, per DOE source). The largest site in Nevada targets 200 MW, enough to power roughly 2-4 large AI training clusters.

The cost reality
Geothermal remains expensive: levelized cost of energy (LCOE) at $60-100/MWh vs $30-50/MWh for solar+storage, per Lazard 2025. But for AI operators paying $100-150/MWh for grid power in data-center-heavy regions like Northern Virginia, geothermal is already competitive. DOE's $165 million investment is modest — less than the cost of a single 100 MW gas turbine plant — but the industry matching brings total to $380 million, enough to de-risk drilling for the first wave.

Unique take: geothermal solves the permitting bottleneck
The real constraint on AI data center buildout isn't compute — it's power interconnection wait times, which PJM data shows average 4 years post-approval. Geothermal plants can be built on-site or within 1-2 miles of data centers, bypassing long-distance transmission lines. The DOE's choice to fund pilots across 8 states signals a deliberate strategy to prove geothermal can skip the grid queue entirely.

What to watch

Watch for the first pilot to reach commercial operation, expected Q3 2027 per DOE timeline. If geothermal LCOE drops below $50/MWh by 2028, expect hyperscalers (Google, Amazon, Microsoft) to announce dedicated geothermal-powered data center campuses, bypassing grid interconnection delays.


Sources cited in this article

  1. DOE
  2. Lazard
  3. Energy
Source: gentic.news · · author= · citation.json

AI-assisted reporting. Generated by gentic.news from 3 verified sources, fact-checked against the Living Graph of 4,300+ entities. Edited by Ala SMITH.

Following this story?

Get a weekly digest with AI predictions, trends, and analysis — free.

AI Analysis

The DOE's geothermal push is smart but small. $380 million total investment is roughly the cost of one 100 MW gas plant. The 5 GW target (by 2030) represents only ~14% of projected AI data center load (35 GW). This is a hedge, not a solution. Crucially, the pilots bypass the transmission interconnection bottleneck that has become the single largest constraint on AI data center buildout. By co-locating geothermal with data centers, the DOE is essentially testing whether firm clean power can be delivered without waiting 4+ years for PJM or MISO queue processing. Google's $5B Texas data center for Anthropic (500 MW by 2026) shows hyperscalers are already betting on gas+renewables mix. Geothermal could disrupt that calculus if LCOE falls below $50/MWh. But the technology remains unproven at scale — Enhanced Geothermal Systems (EGS) have a 30-year history of failed pilots. The 8 new sites will need to demonstrate sustained 90%+ uptime before hyperscalers commit.
Compare side-by-side
Geothermal vs Solar

Mentioned in this article

Enjoyed this article?
Share:

AI Toolslive

Five one-click lenses on this article. Cached for 24h.

Pick a tool above to generate an instant lens on this article.

Related Articles

From the lab

The framework underneath this story

Every article on this site sits on top of one engine and one framework — both built by the lab.

More in Policy & Ethics

View all