Google’s 396 MW Geothermal Bet Sends Fervo Soaring—Can It Arrive by 2028?

The project's timeline is ambitious, with deliveries starting in the third quarter of 2028, but there is no specific information on the current progress or potential obstacles. Geothermal projects often face delays due to permitting, financing, or drilling challenges, which could impact the on-time delivery of the four 99 MW blocks.

Fervo is the strongest enhanced-geothermal (EGS) executor in the market and is well capitalized after its ~$2.2B May 2026 IPO, but the specific proposition here — four successive 99 MW blocks entering service on their target commercial operation dates beginning Q3 2028 without cost overruns — requires several things that were not yet secured as of 1 Sep 2026. The 8-K filed 26 Aug 2026 describes a 15-year PPA between Cape Generating Station 6 LLC and Google Energy LLC for 396 MW delivered in four successive 99 MW sub-tranches with target CODs starting in Q3 2028, plus an offered ~600 MW expansion option. That volume is incremental to Cape Station Phase I (~100 MW, three 33 MW GeoBlocks) and Phase II (~400 MW, eight 50 MW GeoBlocks), of which ~384 MW is already contracted to Southern California Edison and Clean Power Alliance. So the Google tranche means a fresh ~80-well drilling program, new power blocks, and new delivery rights inside roughly 24 months from signature — a tighter runway than Phase II, which broke ground in Q1 2026 for a 2028 COD. The binding constraint is delivery, not resource: Fervo holds only ~290 MW of interconnection and transmission rights at Cape Phase II against 384 MW already contracted, management guided to 2027 curtailment on the Cape lines, and the Google volume leans on direct-to-load contracting under Utah SB132 that Fervo itself describes as subject to regulatory approvals. The load side is equally conditional — the Google data center is expressly 'potential,' subject to engineering feasibility and state and local approvals, with no site publicly announced as of the cutoff. Cost is part of the test too: Enverus (30 Jun 2026) pegs Cape Phase I at roughly $7M/MW against a $5.5M/MW Phase II target and flags unproven sub-1% water loss and thermal drawdown assumptions that could force infill drilling. Fervo had not yet produced first power anywhere at Cape as of the cutoff (GeoBlock 1 targeted Q4 2026, GeoBlocks 2-3 early 2027, versus an original '100 MW to the grid in 2026'). Requiring all four blocks to land on schedule and on budget through roughly 2029-2030 is a demanding conjunction, and the phased sub-tranche structure itself reads as a deliberate schedule shock absorber. I therefore expect at least one of permitting/interconnection, the conditional data center load, or the drilling and cost curve to push or shrink delivery: DELAYED_OR_SMALLER. Note separately that even in an on-time world this is 15 years of contracted output serving a single hyperscaler's new load under a structure designed to insulate other ratepayers, so Utah household bills would barely move — the on-time branch would be 'Google First,' not 'Broad Benefit.'

I select DELAYED_OR_SMALLER. Fervo's rapid drilling progress, strong cash position, and binding Google contract make substantial 2028 delivery plausible, but the stricter outcome—all four 99 MW tranches starting as targeted without expensive slippage—is not my base case. As of the cutoff, Fervo had only a 3 MW pilot operating, its first utility-scale Cape units were still being commissioned, and the Google delivery pathway still required regulatory approvals. Fervo also disclosed substantial additional financing needs and unresolved permitting and delivery constraints. I therefore expect at least one tranche to be delayed, resized, or supported with higher-cost capital. If delivery is timely, the immediate economic benefit will chiefly accrue to Google and the local project area because Utah law assigns large-load incremental costs to the customer rather than existing ratepayers.

As of the 2026-09-01 cutoff, Fervo's Cape Station already has 500 MW under construction, with first power targeted for late 2026, and the Google PPA start target of Q3 2028 leaves roughly two years to ramp the four 99 MW blocks. Project capital is substantially in place via the $421 million Cape Station project financing, earlier Series E/IPO proceeds, and the Google anchor contract; the 396 MW also fits within the 500 MW already permitted. That makes on-time delivery the most likely path. I select ON_TIME_GOOGLE_FIRST because the contracted output is aimed at Google's data-center load, so the project should work without meaningfully moving household bills.

Fervo Energy's 396 MW agreement with Google for its Utah Cape Station project is structured in four 99 MW sub-tranches with target deliveries starting in Q3 2028. Driven by rapid EGS drilling advancements and flexible power delivery pathways authorized under Utah Senate Bill 132 (SB 132), the project is on track to come online as scheduled. Because the electricity is directly committed to Google's proposed Utah data center under a dedicated corporate PPA explicitly structured at no extra cost to existing ratepayers, the newly generated firm clean power directly matches Google's hyperscale computing demand rather than entering general utility rate bases, leaving retail household power bills essentially unchanged.

Fervo’s Cape Station (Beaver County, Utah) is a large next-generation geothermal build, not a completed plant. Google’s 15-year offtake is for 396 MW in four ~99 MW blocks with first deliveries targeted in Q3 2028 and a conditional extra ~600 MW. That timeline is still years away and depends on drilling, wellfield performance, interconnection, permitting, and financing at a scale far above Fervo’s earlier Nevada demonstration. Industry geothermal and EGS projects routinely slip; Cape Station’s own public record already shows phased construction, transmission work, and multi-year well campaigns rather than a locked commercial-operation date. Household-bill relief is not the contract’s purpose—the power is contracted to Google. The most defensible independent position is therefore DELAYED_OR_SMALLER: some capacity may still come online, but delivering all four 99 MW blocks on the advertised Q3 2028 start without costly delay is not the base case.

The Google PPA (SEC 8-K, signed Aug 26 2026) covers four successive 99 MW sub-tranches at Fervo's Cape Station with target CODs beginning Q3 2028. Execution evidence favors on-time delivery: Phase I (~100 MW) is approaching commissioning, Phase II ORC equipment is already contracted (Turboden 180 MWe, Baker Hughes 300 MWe), the site holds BLM permitting approval for up to 2 GW, drilling is ahead of schedule with steep learning-curve gains, and equity, bridge, and project debt financing are secured. Residual EGS scale-up risk (a modest Phase I slip, wellfield scaling, sparse public production data) keeps a delay scenario credible but not dominant. If delivered on time, the power is dedicated to Google's data center load under the PPA, so regional household bills barely move.

As of September 1, 2026, Fervo Energy’s Cape Station geothermal project in Utah is structured in multiple phases, with the first phase (100 MW) explicitly on track to begin delivering power to the grid by late 2026. The full 396 MW capacity, which aligns with Google’s 15-year power purchase agreement, is targeted for delivery starting in the third quarter of 2028. Key evidence supports the likelihood of on-time delivery for the initial 100 MW block, with construction underway, permits secured, and full financing in place. However, the broader 396 MW delivery (comprising four 99 MW blocks) is not explicitly broken out in contemporaneous sources, and the 2028 target for full capacity is subject to risks typical of large-scale geothermal projects, including drilling performance, permitting for expansion phases, and grid integration. The project’s proximity to existing transmission infrastructure and the use of proven oil-and-gas drilling techniques mitigate some technical risks. However, the history of geothermal projects in Utah and the Western U.S. shows that delays in permitting, drilling, or financing can reduce delivery or push timelines. The Western grid is under pressure from rising demand, particularly from data centers, but the Cape Station project is not yet cited as a critical near-term solution for regional power pressure. Given the evidence, the most likely outcome is that the project will deliver its initial 100 MW on time in 2026, with the remaining 296 MW (three additional 99 MW blocks) facing moderate risk of delay or reduction in scale. The primary beneficiary of the initial capacity will be Google, with limited immediate impact on regional power prices or household bills.

I select ON_TIME_GOOGLE_FIRST. The evidence supports delivery by 2028 more than a major delay: Phase II construction had begun, initial wells were drilled, long-lead equipment was secured, a third drilling rig was added, and Fervo reported a 21-day record for its ninth 3.0 well. The $2.2 billion IPO also materially strengthens near-term liquidity. However, this remains a first-of-kind, technically complex eight-GeoBlock expansion requiring extensive drilling and construction, while DOE’s environmental review explicitly was not a final loan-guarantee or construction approval and Fervo remains responsible for permits. Thus on-time delivery is plausible but not sufficiently certain for a high-confidence call, and the contracted power is primarily tied to Google’s prospective Utah data-center demand rather than broad household bill relief. [web:29][web:41][web:46]

Fervo Energy's Utah project relies on next-generation Enhanced Geothermal Systems (EGS) using horizontal drilling and multi-stage fracturing, technologies adapted from the oil and gas industry but unproven at commercial utility scale. While Google has secured a 15-year agreement for 396 MW targeting Q3 2028 delivery, historical precedents for novel energy infrastructure indicate significant risks related to subsurface uncertainty, permitting delays, and supply chain constraints for specialized drilling equipment. The complexity of scaling four 99 MW blocks simultaneously increases the probability of technical bottlenecks or phased delays rather than a uniform on-time commissioning. Therefore, the risk of delays or reduced initial capacity is higher than the likelihood of flawless, on-time full-scale delivery.

As of September 2026, enhanced geothermal projects in the U.S. have consistently faced well-documented delays from permitting backlogs, exploratory drilling risks, and complex project financing closures, making on-time delivery of four 99 MW Fervo Utah blocks unlikely by Q3 2028.

Fervo's Cape Station Phase II (400 MW, eight 50 MW GeoBlocks) is targeted for 2028, backed by a completed DOE environmental review (July 2026), ~$2.2B IPO proceeds, $421M project debt, and record drilling. The 396 MW Google PPA maps to this phase with Q3 2028 deliveries. But the power is a dedicated, largely behind-the-meter offtake for Google's Utah data center ('no cost to existing ratepayers'), so it eases regional household bills only marginally. Execution and at-scale EGS risk keep delay probability material but below on-time.

As of the September 2026 cutoff, Fervo's Cape Station buildout in Utah has shown strong execution—Phase 1 drilling came in faster and cheaper than planned, and the project secured roughly $678 million of project financing in late 2024—so the Google 396 MW expansion (four 99 MW blocks, targeted deliveries from Q3 2028) is a credible, well-funded project. However, the question requires ALL four blocks to reach commercial service on schedule with no costly overruns, a high bar for first-of-a-kind enhanced geothermal at scale: each block depends on new well pads in a still-maturing reservoir, plus turbine supply, PacifiCorp interconnection, and roughly two years of remaining construction exposure. Megaprojects of this type routinely slip by quarters rather than deliver clean conjunctions, and any drilling setback or equipment delay would push at least one block past Q3 2028 or trim delivered capacity. Even if on time, the power is contracted to a single corporate offtaker, so it mainly serves Google's data-center load growth rather than materially lowering regional household bills. I therefore expect at least one block to be delayed or the total delivered scope reduced, with modest probability of full on-time delivery.

Fervo's Utah geothermal project faces significant permitting, financing, and technical risks that make on-time delivery of four 99 MW blocks by Q3 2028 unlikely; while Phase I has financing, Phase II lacks full funding, EGS is early-stage, and geothermal projects typically take 7-10 years from lease to operation.