U.S. geothermal developer Fervo Energy said in its third-quarter 2026 results, released on November 9, 2026, that it has signed a 15-year agreement to supply 396MW of power to Google. As a result, total future revenue under binding power purchase agreements rose to $11.9 billion, up from $7.2 billion at the end of the second quarter. However, the $11.9 billion is the combined future revenue from multiple contracts. It does not represent quarterly sales or current cash on hand. For Fervo, which uses horizontal wells to extract underground heat for power generation, securing long-term power sales contracts and building plants that actually deliver electricity are separate stages that should be evaluated separately.
The new contract gives Fervo secured demand for large-scale geothermal power, but contracted capacity is not the same as capacity that is actually operating. The company plans to add facilities using longer horizontal wells while also cutting construction costs per kW of generating capacity. Key factors for judging the business's growth potential will be whether it can deliver the contracted power on schedule and how far it can bring down construction costs as it expands.
A firm 396MW contract and an option for about 600MW more
The new agreement with Google covers 396MW of power delivered through the grid over a 15-year term. The contract uses a "take-or-pay" structure, under which payment obligations arise regardless of how much power is actually purchased. However, the November 9 results announcement did not disclose how plant outages are treated or the specific billing terms.
With this contract, the supply capacity Fervo has secured under binding agreements increased from 658MW at the end of the second quarter to 1,054MW. The corresponding total contracted revenue grew from $7.2 billion to $11.9 billion. Because the $11.9 billion includes multiple contracts, however, it is not possible to calculate a per-unit electricity price using only Google's 396MW and the 15-year term.
The collaboration with Google also includes an option to purchase about 600MW of additional capacity and a development framework of up to 3GW, in addition to the 396MW now confirmed. These differ in how firmly they are committed under contract, so it would not be appropriate to add them all together as confirmed supply capacity. Based on Fervo's announcement, the position of each arrangement is as follows.
| Arrangement | Capacity | Contractual status | Term / delivery format |
|---|---|---|---|
| Power purchase agreement with Google | 396MW | Binding contract signed | 15 years, delivered through the grid |
| Google's additional purchase option | About 600MW | Conditional on Google exercising the option | Direct supply to customers under consideration for the added capacity |
| Development framework with Google | Up to 3GW | Purchase of the full capacity is not confirmed | Framework for development through 2033 |
The development framework of up to 3GW was agreed in March 2026 and differs in nature from the firm 396MW contract. If Google exercises its option for about 600MW, contracted capacity combined with the 396MW would reach about 1GW, but the purchase of the additional capacity is not confirmed at this point. In addition, while the 396MW is planned to be delivered through the grid, a different delivery format is being considered for the roughly 600MW of added capacity. Fervo is considering supplying power directly to customers under a framework based on Utah's SB132, and it lists supplying a data center that may be built in Beaver County as one option.
Using longer horizontal wells to speed up Cape Station expansion
To fulfill its power sales contracts, Fervo needs to expand both the wells that extract underground heat and the surface generation facilities. The company uses horizontal drilling and monitors facility performance with fiber-optic cables installed in the wells to measure temperature, flow rate and other data. Geothermal power is seen as a complement to weather-dependent solar and wind, but that does not mean individual plants can run continuously without interruption.
Fervo and Google's collaboration also already has a track record of commercial operation. In November 2023, Google announced that an early geothermal project in Nevada had come online and begun delivering power to the local grid. That project used two horizontal wells and underground measurement technology based on fiber optics. The new 396MW contract is separate from that early project and expands Cape Station, the large-scale geothermal plant Fervo is developing in Utah.
The expansion plan for Google is called "Cape Google 400" and corresponds to Cape Station's units 4 through 11. As of the November 9 announcement, drilling was about 18% complete, and the first 50MW turbine group had passed factory acceptance testing. Fervo plans first power delivery in mid-2028 and aims to have the full capacity operating by the end of that year.
The expansion will use the "3.0 design," which employs longer horizontal wells than before. The plan is to raise the output per well by extending the horizontal section, enlarging the diameter of the casing that protects the well, and drilling deeper. A longer horizontal section increases the stretch of well in contact with underground heat, but how much output actually improves as a result will need to be verified with real operating data.
Sawtooth 7, the ninth well of the 3.0 design, which Fervo announced in July, reached a horizontal section length of 7,500 feet. That is 50% longer than the 5,000 feet of the earlier 2.0 design. The casing diameter was also increased from the previous 7 inches to 8 5/8 inches. The time from the start of drilling to reaching a total drilled length of 19,448 feet was 21 days. However, that is the drilling time for a single well and does not represent the construction period for an entire plant.
Meanwhile, Cape Station's earlier units are moving toward commercial operation in stages. Unit 1 began commercial operation at 33MW on September 30, 2026. Unit 2 is progressing through grid synchronization and test power delivery, and Unit 3 is at the stage of mechanical completion and cold commissioning. Fervo plans for the remaining two units to begin commercial operation by January 1, 2027. Therefore, not all of the roughly 100MW of initial development capacity is already in commercial operation.
Construction cost targets must be distinguished from actual costs
To turn large power sales contracts into earnings, Fervo needs not only to complete its plants on schedule but also to keep construction costs down. The company aims to cut costs through improvements to its horizontal wells and standardization of facilities, but the figures currently published mix estimated costs for the initial units with future reduction targets. When comparing them, it is important to pay attention to the scope of facilities covered and to the differences between actual results, estimates and targets.
| Construction / installation cost | Scope | Status of the figure |
|---|---|---|
| About $7,000/kW | Cape Station Units 1–3 | Estimated total cost |
| $5,500/kW | Cape Google 400 | Installation cost target |
| $3,000/kW | Future power plants | Long-term installation cost target |
There is a gap of $1,500 per kW between the estimated cost of the initial units and the target for the Google expansion plan. However, $5,500/kW is not a confirmed actual construction cost for completed facilities. The same applies to the long-term target of $3,000/kW: whether cost reductions have been achieved cannot be judged without checking actual costs after the facilities are completed.
Nor can the amount capitalized on the balance sheet be treated as the construction cost of an individual unit. With Unit 1's start of commercial operation, Fervo transferred $514.3 million from construction in progress to property, plant and equipment on September 30. That amount includes not only Unit 1 but also the Phase I well field and infrastructure shared by Units 1 through 3, as well as capitalized interest and financing costs.
Fervo also explains that this $514.3 million is not the construction cost of Unit 1 alone. If the amount were divided by Unit 1's 33MW alone, the calculation would load the cost of infrastructure shared across multiple units onto Unit 1. To compare actual construction costs per kW of capacity, the treatment of shared facilities and the scope covered would therefore need to be aligned. Once the breakdown of capitalized amounts becomes clear with the completion of subsequent units, the cost structure of the initial units will be easier to evaluate properly.
The timing of capital expenditure should also be distinguished from total construction cost. Fervo lowered its capital expenditure outlook for the second half of 2026 from $850–900 million to $700–750 million. However, this was because some spending moved to a later period as construction schedules were adjusted, and the company has not changed its plant operating schedule or its outlook for total investment. Therefore, the decline in second-half capital expenditure alone cannot be taken as evidence that it has succeeded in cutting construction costs.
Next focus: 2028 power delivery and the cost of expansion
Fervo also reports that the development potential of Cape Station exceeds earlier assessments for its underground resources. In an assessment of the initial underground heat in place, commissioned by the company and carried out by DeGolyer and MacNaughton, Cape Station's potential generating capacity was raised from the previous 4.3GW to 8.4GW. The result reflects the 3.0 design, which enables deeper drilling, and the acquisition of additional leases.
However, the 8.4GW figure is a potential capacity estimated from the thermal energy present underground before development, not the capacity of power plants already operating. It also differs from capacity for which buyers have been secured through power purchase agreements. How much generating capability can actually be developed depends on how underground temperature and flow rate change over the long term, as well as on well drilling costs and surface facility construction costs.
On the financial side, cash and cash equivalents as of the end of September 2026 were $1.9 billion, and capital expenditure in the third quarter was $256.7 million. The operating loss for the quarter was $43.7 million, a figure of a different nature from the $11.9 billion in future contracted revenue. While long-term power sales contracts support the outlook for future earnings, actual revenue requires completing facilities and bringing them into commercial operation, and the investment needed for that must continue.
The immediate points to watch are whether the remaining two initial Cape Station units can begin commercial operation by early 2027, and whether first power delivery to Google and full-capacity operation can be achieved in 2028 as planned. Another key question is how close the actual construction cost of the completed expansion facilities will come to the $5,500/kW target. If the output gains from longer horizontal wells and the cost reductions from repeated construction are demonstrated, it will become clearer at what cost Fervo can deliver stable supply against the large power demand it has secured.