DOE Picked 21 Geothermal Projects and Funded Only the First Budget Period. The Rolling NOFO Is the Real Opportunity.

Two numbers from the Department of Energy's September 21 geothermal announcement do not match, and the gap between them is the most useful thing in the release.

DOE selected 21 projects for up to $99 million. The underlying notice of funding opportunity, DE-FOA-0003472, was issued February 25, 2026 with up to $171.5 million available. The $99 million covers the first budget period only. Everything after that is, in DOE's own language, "subject to Congressional appropriations and successful completion of agreed-upon milestones and performance criteria."

That is not a shortfall. It is a structure — and once you see it, the FY2027 version of this competition looks very different from a one-shot deadline race.

What Got Selected

The portfolio splits cleanly along the solicitation's two funded topic areas:

• Five projects for enhanced geothermal systems field tests. The field-test topic carried up to $100 million and anticipated 4 to 10 awards. Five selections sits at the bottom of that band.
• Sixteen projects for exploration drilling to characterize and confirm promising next-generation and conventional hydrothermal prospects. That topic carried up to $71.5 million and anticipated 8 to 18 awards. Sixteen sits near the top.

The program office is the Hydrocarbons and Geothermal Energy Office (HGEO) — the consolidated office that now carries geothermal alongside conventional subsurface energy, and the same office running the University Training and Research program whose letters of intent are due October 1. Under Secretary Kyle Haustveit framed the selections as empowering "American innovators to unlock the tremendous geothermal resources beneath our feet," emphasizing around-the-clock power and energy security rather than emissions.

Publicly surfaced selectees sketch the portfolio's shape without DOE having released per-project federal amounts: the University of Utah picked up two projects; a Fort Worth-based energy company will drill in Oregon; a Salt Lake City company building AI-driven subsurface tools has a New Mexico project; an Arlington, Virginia energy consulting firm has a California project. Data from all 21 projects will flow into DOE's Geothermal Data Repository, which is a competitive fact rather than a courtesy — every well logged under this program becomes free input for the next applicant's resource case.

Read the Award Ratio Before You Read the Technology

Divide selections by topic-area ceilings and the two tracks tell opposite stories about competitive difficulty.

The exploration-drilling track awarded 16 projects against a $71.5 million ceiling — an average in the low single-digit millions per project. That is a track designed to buy many wells at modest federal exposure, because the point is de-risking: confirm whether a resource exists before anyone finances a power plant on it. Smaller developers, state geological surveys, and university-industry teams can compete in that pool with a well-scoped single-well program.

The field-test track put up to $100 million behind five projects. Even before negotiation, that implies federal shares in the eight figures for projects that must demonstrate stimulation, flow, and sustained production at field scale. This is not a track a first-time applicant wins on narrative quality. It rewards teams that already control acreage, already have subsurface data, and can carry a serious match.

The strategic error we see repeatedly in DOE energy competitions is a team choosing the topic area with the bigger headline number. The correct move is choosing the topic area whose average award matches the project you can actually execute and finance. Fifteen million dollars of federal money on a project your balance sheet cannot match is a loss dressed as ambition — the same lesson the SPARK transmission selections taught three days later, where recipients brought $3.35 billion against $1.9 billion federal.

The Rolling NOFO Changes the Calendar Math

Here is the provision most applicants missed. DE-FOA-0003472 is structured as a rolling solicitation: it may reopen for up to 72 months, with review cycles approximately every 12 months, contingent on appropriations. As of late September 2026 the NOFO is closed.

Three consequences follow.

Missing the May 14 deadline cost you a year, not the program. The FY2026 cycle ran letters of intent March 27, full applications May 14, selections September 21 — roughly six months from LOI to selection. If the cycle repeats on that cadence, the next LOI window plausibly falls in the first quarter of 2027. Teams that were not ready in March 2026 have a defined re-entry point instead of an indefinite wait.

The first cohort's data becomes your evidence base. Twenty-one projects feeding the Geothermal Data Repository means the next applicant in the Great Basin, the Cascades, or the Permian's geothermal margins can cite public characterization data rather than commissioning it. In a resource-confirmation competition, borrowed subsurface data is worth more than borrowed prose.

Reviewers will have seen the first cohort's scopes. A rolling NOFO reviewed by a stable program team develops institutional memory. Cycle-two proposals that read as thinner versions of funded cycle-one projects fare badly; proposals that address a gap the first cohort left — a basin nobody drilled, a stimulation approach nobody tested, an offtake structure nobody demonstrated — read as portfolio completion.

Selection Is Not an Award

DOE's boilerplate deserves quoting because teams keep treating it as boilerplate: "Selection for award negotiations is not a commitment by DOE to issue an award or provide funding."

What happens between selection and executed agreement is where geothermal projects in particular die. Negotiation pins down scope, budget, milestones, and cost share. For a drilling project, that means the rig contract, the permit status, the surface-use agreement, and the water source all get examined as commitments rather than intentions. A rig slot that was "available" in May at a quoted day rate is a different conversation in November. Federal land access adds a Bureau of Land Management review that no cooperative agreement can accelerate.

If you are on the September 21 list, the operative deadlines for the next 90 days are yours to set: convert conditional cost-share letters into board-approved commitments, lock drilling contractor availability against a milestone schedule you can defend, and confirm every permit assumption in the application still holds. The budget-period structure adds a second discipline — the first period's milestones are the gate to continuation funding, so negotiate milestones you can hit on a schedule that survives one bad weather season or one rig delay. Milestones written optimistically to look aggressive in negotiation become the instrument of your own de-scoping in year two.

Why Geothermal Keeps Getting Funded

The portfolio context matters for anyone deciding where to spend proposal effort in FY2027. Geothermal has been the durable exception in federal clean-energy funding through 2026 — we flagged the pattern in May when DOE committed $171.5 million to field tests plus $30 million to superhot rock drilling through ARPA-E while solar and wind tax credits were phasing down.

The reasons are structural rather than rhetorical. Geothermal is firm, dispatchable, around-the-clock generation, which fits a grid narrative dominated by data-center load growth. Its workforce, supply chain, and technical vocabulary are the oil and gas industry's — directional drilling, well completion, reservoir stimulation — which makes it legible to an administration whose energy policy centers domestic hydrocarbons, and which is precisely why the program office is now the Hydrocarbons and Geothermal Energy Office. And it enjoys genuine bipartisan permitting support, with multiple bills advancing to bring geothermal leasing closer to the categorical treatment oil and gas already receives.

For applicants, that adds up to a program with unusual appropriations resilience. The $99 million first-budget-period structure is a hedge against a continuing resolution, not a signal of retreat.

Three Moves for the Next Cycle

Pick the topic area by average award, then build the match to it. Exploration drilling for teams that need a well confirmed; field tests only if you can carry eight-figure project economics.

Start the LOI in December, not March. A six-month LOI-to-selection cadence implies a first-quarter 2027 window. Site control, permits, and a rig letter take longer than the writing does.

Mine the repository before you write. Twenty-one newly funded projects will publish characterization data. A cycle-two proposal that cites cycle-one public data to de-risk its own resource case is arguing with evidence the reviewers helped create.

Tracking which DOE solicitations are rolling, which are one-shot, and when each one plausibly reopens is the part Granted was built to shorten.