For 8.7 seconds, the Adélie prototype reportedly produced more electricity than it consumed at 1,500 metres beneath the Pacific. That is the result announced by Endurance Energy in a press release dated 19 October 2026, concerning a test conducted on 27 September at Axial Seamount volcano, off the coast of Oregon. Futura revisits the experiment on 12 November. The question is what this interval measures, and what steps still separate a prototype from a continuous power supply.
The figure sums up a specific question: can a geothermal installation operate at great depth and supply more energy than it uses? Endurance’s answer is yes, for a short period. The challenge remains turning this demonstration into sustained operation in an environment where fluid flow, equipment and maintenance operations impose particular constraints.
What “net positive” means
Adélie is a generator weighing about three tonnes, installed on the seabed at a depth of around 1,500 metres, on Axial Seamount volcano, about 480 kilometres off the coast of Oregon. To generate electricity, the system uses the heat of a hydrothermal fluid: this heats a working fluid kept in a closed circuit, which drives a turbine and an alternator.
“Net” generation is the electricity produced after deducting the electricity consumed by the system to operate. Endurance’s press release reports that the balance was positive for 8.7 seconds. Over that period, median net power was about 2 kilowatts, with a peak of 16.5 kilowatts.
These figures describe different things. The median is the midpoint of the measurements, not their average over a long period. The peak indicates the maximum power reached, not a stable output. A kilowatt measures power—that is, the rate of energy flow at a given moment—while a kilowatt-hour measures an amount of energy accumulated over time. The figures available therefore characterise a brief performance, not the daily or annual output of a future power plant.
The well and the vent are not the same heat source
The test involved two operations: drilling a geothermal well on the seabed and running Adélie using hot fluid. The distinction is crucial. The well drilled by Endurance reaches 16 metres and produced fluid at 260 °C after it was completed. But its flow rate was too restrictive to supply the power plant during the test.
The net-positive period was achieved using a nearby natural hydrothermal vent, whose fluid was at about 281 °C. In other words, the drilled well did encounter heat, but it was not the source of the flow used in the successful electricity test. Flow rate, not just temperature, is essential: enough heat must be supplied to the system to generate electricity continuously.
This difference highlights the main challenge identified for the next phase. Endurance believes deeper wells could provide a higher flow rate. This remains a technical possibility to be tested, not a result already achieved with Adélie.
An interrupted run, followed by a second failure
According to the company’s account, a battery overcurrent protection system stopped the first net-positive run. About ten minutes later, during an attempt to restart, an electrical cable supplying the working-fluid pump failed, bringing the dive to an end.
These incidents interrupted the test; they do not change the meaning of the reported measurement, but they show what the brief run has yet to establish: the ability to restart, keep equipment in service and generate power for a long time under real-world conditions. Endurance describes Adélie as a test prototype, not a commercial-scale power plant. The company’s announcement is not a peer-reviewed study.
The press release also describes a three-week campaign involving drilling, deployment and remote operation, communications tests, and the circulation of hot fluid. This duration refers to the deployment, not three weeks of continuous electricity generation. In its 19 October press release, Endurance says it had sent more than 40 water, rock and mineral samples to independent laboratories for analysis. The document does not publish their findings.
Heat from below the seabed, ecosystems on the seafloor
Hydrothermal vents form when seawater enters cracks in the ocean floor, heats up in contact with rocks warmed deep underground, and rises carrying chemical substances. These fluids sustain distinctive biological communities: microbes use their chemical compounds and, in turn, form the basis of food webs that include, among others, tube worms, crabs and molluscs.
Endurance says it mapped the seafloor before installing the equipment, took water samples before and after operations, and monitored the sites on video. The company reports that it detected neither a major change in the seabed community nor a significant change in the site’s hydrothermal activity. This is an account presented by the operator based on its monitoring programme; it is not enough to establish the effects of prolonged operations or a larger-scale deployment. The announced sample results may provide further evidence.
The next steps announced
Endurance plans to deploy Gentoo during the winter, an upgraded version of Adélie intended to target higher power and more sustained generation. The company is also developing Leopard, a commercial generator announced at 2 megawatts. This power rating is a design target: it was not achieved by Adélie.
A first commercial project is under development in Tonga, with initial production targeted for 2027. Here too, this is a date announced by the company, not an installation already in service. Between 8.7 seconds of positive electrical balance and a sustainable power supply, components still need to be made more reliable, sufficient flow secured, deep-water interventions organised and environmental effects assessed. The test has made these questions more concrete; it has not resolved them.
References
- Futura, overview of the Adélie test
- Endurance Energy, press release on Adélie
- Woods Hole Oceanographic Institution, how hydrothermal vents work
The El Gordo hydrothermal site at Axial Seamount, with oceanographic observation equipment. Archive photograph: it does not show the Adélie prototype or its 2026 test.
UW/NSF-OOI/CSSF — Ocean Observatories Initiative — Wikimedia Commons — Public domain.
