The power shortage in the United States is not a new issue for iSuga, but the situation continues to worsen. The particular problem lies in the fact that electricity demand from data centers operating AI services has surged far beyond initial expectations.
Goldman Sachs estimates that U.S. data center power demand will surge from 31 gigawatts (GW) last year to 41 GW this year and 66 GW next year. The 10 GW increase this year is equivalent to the output of 10 nuclear reactors, while the 25th GW increase next year matches the capacity of 25 nuclear reactors. The U.S. Department of Energy has warned that without an additional 100 GW of power by 2030, annual blackout duration will jump from eight hours to 800 hours—a 100-fold increase. That 100 GW is comparable to the total electricity consumed across all of South Korea during midsummer. The current reality is that any source capable of generating electricity—from nuclear power to desert solar farms and offshore wind—is being pursued without discrimination.
The problem is that it takes considerable time to install and operate most power generation facilities, and there are also significant drawbacks. Solar power cannot operate on cloudy days or at night, and wind power relies entirely on the wind. Gas power plants emit carbon, and small modular reactors (SMRs) will not be available until at least 2030. This is why technologies such as hydrogen fuel cells, once considered unprofitable, are now gaining prominence. Perbo Energy holds technology for generating electricity using geothermal energy and has recently attracted attention from Wall Street.
Perbo Energy is a company founded by Tim Latimer, a young engineer who was drilling wells in oil fields. After graduating from university, Latimer worked as a drilling engineer for Australian mining firm BHP in 2012. His workplace was Permian and Eagle Ford, the central hubs of U.S. shale gas development. While working there, he acquired various drilling technology know-how. In the past, wells were drilled vertically or horizontally by laying them on their sides, or rock formations were fractured using water. The longer the drilling section, the higher the underground temperature becomes; if this is not properly managed, water that seeps in can instantly vaporize and trigger an explosion.
Latimer, while pondering the geothermal issue, one day conceived a plan to utilize geothermal energy instead of oil. In fact, geothermal power generation is not a new technology, but commercializing it is a different matter. For geothermal power generation to be feasible, one must find land with cracks that allow water to seep through. Existing geothermal power plants could only operate in places where hot water naturally rises, like hot springs. This is why they are limited to volcanic regions such as Iceland or California's Geysers.
Latimer decided to apply his experience in shale gas drilling to research water injection technology for wells, quit BHP, and enrolled in Stanford University's graduate school. While pursuing both an MBA and a master's degree in environmental resources, he met Jack Norbeck, who was researching geothermal reservoirs. This was a combination of a drilling practitioner and a subsurface simulation expert. The two founded Fervo Energy in Houston in May 2017. Investors included Bill Gates' Breakthrough Energy Ventures, shale giant Devon Energy, and Google. Norbeck continues to serve as the chief technology officer (CTO).
The two individuals founded Perbo Energy, which is based on a technology called Enhanced Geothermal Systems (EGS). The name may sound complicated, but its structure resembles Korea's ondol heating system. Below 3 kilometers underground, granite exceeds 200℃. A well is drilled vertically to this depth, then extended horizontally for an additional 2 kilometers. Two such wells are drilled side by side: one is an injection well where water is pumped in, and the other is a production well that extracts hot water. The rock formation between the two wells is fractured dozens of times, similar to shale, to create pathways for water flow.
When cold water is injected, it heats up as it passes through the bedrock and rises to an adjacent well as hot water. This heat drives a turbine. The design has now evolved from using just two wells to a "wine rack" configuration, where multiple wells are stacked like wine racks on shelves. Each well covers a larger volume of bedrock, and the water paths do not overlap. The deepest well at the Cape Station reaches a total depth of 6,054 meters — equivalent to stacking 11 Lotte World Towers.
While conventional geothermal power plants must seek out natural hot springs, Perbo Energy can build a power plant anywhere as long as there is hot rock. This eliminates site constraints. Perbo Energy's power plants use binary generators from Turboden of Italy, a subsidiary of Mitsubishi Heavy Industries. The system works by using hot water to heat a liquid with a low boiling point, which then drives the turbine.
Perbo Energy's competitiveness lies in its ability to simulate and analyze subsurface structures. This method involves inserting fiber-optic cables into boreholes to analyze the sounds of rock fracturing and the temperatures of flowing water, utilizing distributed acoustic sensing (DAS) technology common in the shale industry. It allows operators to see from the surface where blockages occur and where water leaks. Novex's simulations developed at Stanford University are integrated into this system.
Although there are many geothermal power companies, Perbo Energy has attracted attention because its drilling data is reliable. The first well in Nevada was completed in 70 days in 2022, but at the Utah Cape Station in 2024, the period was reduced to just 21 days. A significant portion of geothermal power plant costs goes toward drilling wells, and Perbo Energy boasts a very high success rate. The company initially estimated its success rate at 18%, but the actual result reached 35%. In June 2025, it drilled a well with a depth of 4,805 meters and a temperature of 271°C in just 16 days. In July this year, it completed a well with a total length of 5,928 meters, a horizontal section of 2,286 meters, and a temperature of 238°C in 21 days. The drilling speed has become 143% faster than at the first Cape site.
Power generation performance is also improving. The first generation produced 3 MW from a single production well. The second generation reached 10 MW, and the third generation currently being drilled achieves 15 MW. This is because the diameter of the drill hole has more than doubled from 12 cm, its length has increased from 914 m to 2,286 m, and the geothermal source powering power generation has risen from 177°C to 221°C. Equipment investment costs were at a level of 20.25 million won (15,000 dollars) per kW, which dropped to 9.45 million won (7,000 dollars), and have now improved to a level of 7.43 million won (5,500 dollars). The industry views the point where horizontal sections reach 4,572 m, temperatures exceed 260°C, production wells generate 25 MW each, and costs fall to 4.05 million won (3,000 dollars) per kW as the threshold for achieving economies of scale.
Perbo Energy is currently operating three Helmerich & Payne (H&P) drilling rigs. Each rig can create 1.5 wells per month, or 54 wells annually. This translates to a construction speed of over 400 megawatts per year. Third-generation wells generate 27% more electricity at the same cost compared to previous generations. The technology has also been fully validated. Perbo Energy's Nevada "Project Red," which served as a pilot site for data centers, operated for over 600 days after commencing in 2023, achieving an availability rate of 98.4%. No repairs were needed on the wells, and the temperature drop over 500 days was limited to just 1.4℃. The Cape Station test well produced more than 10 megawatts from a single production well, three times the output of Project Red.
Perbo Energy signed a massive power supply contract with Google, a major shareholder, on the 1st of this month. The agreement entails supplying 396 megawatts of geothermal electricity extracted from underground in Utah's desert to Google for 15 years. Google has also offered an option to purchase an additional 600 megawatts; combined, this approaches 1 gigawatt. It is the largest contract in U.S. geothermal power history, equivalent to the capacity of one nuclear reactor. On the day the contract was made public, Perbo Energy's stock price rose by more than 30% during trading hours.
In addition to Google, numerous data center companies are paying attention to Power Energy. The reasons are multifaceted, but the most significant factor is the short period required to build power facilities. Google's investment in the SMR company Kairos Power targets its first reactor for 2030, with completion of a 500-megawatt facility scheduled for 2035. In contrast, Power Energy will deliver its first batch in the third quarter of 2028. Latimer stated during the second-quarter earnings announcement that "for companies needing carbon-free baseload power before 2030, Power is practically the only option."
The price is also competitive. The PPA unit price the company is negotiating ranges from 135,000 to 175,500 won per megawatt-hour (MWh), or $100 to $130. This is lower than the new nuclear power generation cost of 190,000 to 298,000 won ($141 to $221) and gas peaking power costs of 155,000 to 298,000 won ($115 to $221), as compiled by investment bank Lazard.
A key advantage is that it requires a smaller land area and achieves higher resource recovery rates compared to solar power. According to an analysis by Wonri Research, the entire Cape Station site for a 2GW facility covers 2.55 square kilometers, which is only one-twentieth the size of a solar power plant with the same capacity. Water circulates underground, achieving an utilization rate of 83%, four times that of solar power (in the 20th% range). Based on standards from the U.S. National Renewable Energy Laboratory (NREL), lifecycle carbon emissions are 37 grams per kilowatt-hour, one-thirteenth that of gas power generation (486g) and one-twenty-seventh that of coal (1001g).
In addition, Perbo Energy also succeeded in experimenting with an "underground battery (FervoFlex)" that pumps water into the ground when electricity is surplus and extracts it when needed. The system completed five cycles of 12-hour charging and discharging. Another merit is that geothermal power generation produces carbon-free electricity, unlike thermal power generation. California mandated in 2021 that power companies secure 1,000 megawatts of carbon-free power sources independent of weather conditions.
Perbo Energy's confirmed PPA stands at 658 MW as of the end of June, with an additional 396 MW from Google now added. The order backlog is 9.72 trillion won ($7.2 billion) as of the end of June. This means a company with no revenue yet has secured 15 years' worth of sales in advance. Google also signed a basic contract in March to purchase up to 3 GW on a priority basis. Among customers currently under negotiation, hyperscalers account for 50%, power companies 35%, and industrial firms 15%. Recently, the company has also been promoting a "behind-the-meter (BTM)" model that places power plants directly next to data centers without using the transmission grid. With no transmission fees involved, as data centers grow larger, they can simply attach "geoblocks" in 50 MW increments alongside them. The first BTM contract is expected to be announced within this year.
Of course, competitors are also moving quickly. Next-generation geothermal power technologies are broadly divided into two types: Perbo Energy and Sage GeoSystems have chosen the method of breaking bedrock to let water flow through it. In contrast, Canada's Eavor and U.S. XGS Energy use a method of burying pipes underground and circulating water only within them, akin to burying radiators in the ground.
The advantage of the IBER method is that it does not fracture bedrock, eliminating earthquake risks and preventing water loss. Meta's reason for contracting with XGS for a 150-megawatt project in New Mexico was also "zero water usage." The drawback is that the heat exchange area is limited to the pipe surface. The first power plant operated by IBER in Germany last December, which drilled four pairs of wells totaling 16 kilometers, produced only 8 megawatts of electricity output. This falls short of the output of a single Perbo Cape well. The German government subsidy rate is approximately 250 euros per megawatt-hour.
Sage uses "pressure geothermal," which, like Perbo, drills up to a depth of 6 kilometers and pushes water out using pressure. Sage has signed a contract with Meta for 150 megawatts and received an investment of 131 billion won (97 million dollars) in January this year, led by Ormat Technologies, the world's leading geothermal company. There are no commercial power plants yet. Quaise Energy is a company that plans to drill up to 20 kilometers by melting bedrock using millimeter waves based on the principle of microwave ovens. Its current drilling record stands at 100 meters.
The established strong player Omat operates 1.3GW of geothermal and solar power globally, but its facilities are natural hot spring-based. It has almost no new sites available. This is why the contract it signed in January this year with data center company Switch was limited to just 13MW. The difference from Perbo Energy lies in commercial viability. Perbo Energy has drilled more than 25 wells, reducing drilling time by 75% and drilling costs per meter by 70%, and it is openly sharing various figures, whereas other companies do not disclose such data.
Perbo Energy has approximately 260 employees, most of whom are in research and development or as field engineers, with a particularly high number coming from the shale industry. A TRIR (Total Recordable Incident Rate) of 0.34 is an excellent level even by oil industry standards. The management team and shareholder structure are also strong. Meg Whitman, former CEO of HP and eBay, serves as an independent director. Board members include Jessica Ull, a former chief financial officer at Shell, and Troy Low, former chief technology officer at Devon Energy. Devon Energy, a major shale company and the largest shareholder, holds a 12.5% stake.
Perbo Energy is particularly leveraging the equipment and know-how of companies specializing in shale oil development. Its financial investors are also robust. Eight banks, including Barclays, JPMorgan, and MUFG, provided 568.9 billion won (421.4 million dollars) in project financing. This is being evaluated as the first major project financing for next-generation geothermal energy backed by the banking sector.
The regrettable point is that Perbo Energy's financial statements are not yet reliable. Last year, revenue was 190 million won, with a net loss of 78 billion won (577.9 million dollars), and in 2024, the net loss was 55.5 billion won (411.1 million dollars). Revenue for the first half of this year was only 2.75 trillion won (174,000 dollars), and the net loss reached 118.4 billion won (877.4 million dollars). However, the company has sufficient cash on hand. Perbo Energy raised 275 billion won (2.04 billion dollars) when it listed on the New York Stock Exchange in May this year, and all of these funds remain intact.
The assets under construction amount to 1.67 trillion won ($1.235 billion), and the planned capital investment for the second half of the year is between 1.15 trillion and 1.215 trillion won ($850 million to $900 million). There will be no major funding issues through the end of 2027. Meanwhile, Cape Station Phase 1, with a capacity of 100 megawatts (three 33-megawatt geoblocks), is scheduled to begin full-scale operations by the end of this year or early next year, which is expected to increase revenue figures. According to Perbo Energy's plans, the current development site covers 2,630 square kilometers, and the pipeline capacity exceeds 50 gigawatts.
Wall Street analysts view Perbo Energy's potential positively, but some caution that it remains to be seen when the virtuous cycle of investment-revenue-profit will be fully realized. The company has set its 2027 revenue guidance at 81 billion to 108 billion won ($60 million to 0.8 billion won). This figure is less than half of what could be generated if a 100-megawatt facility were operating at full capacity. Cape Station's transmission line will now be shared with other power plants in the region, resulting in output curtailment for the entire year of 2027. In effect, there are no transmission lines available to supply the electricity that has been generated. Following this news, the stock price fell by 16% on August 12.
Rising producer prices in the United States and soaring industrial costs could also pose a burden. Perbo Energy is estimated to require construction costs of 95 billion won (7 million dollars) per megawatt, which is more than five times that of solar power. The company also has high dependence on U.S. investment tax credits (ITC). Phase 1 financing includes an ITC bridge loan of 82.3 billion won (61 million dollars). Additionally, the Google contract contains a clause stipulating that if Perbo fails to propose an additional 600-megawatt option, it must pay a "power supply compensation fee."
Nevertheless, Wall Street's overall assessment remains positive. In June, shortly after Perbo Energy's listing, J.P. Morgan issued a "buy" recommendation with a target price of 63,450 won ($47). Piper Sandler set its target at 68,850 won ($51), RBC at 62,100 won ($46), and Baird at 63,450 won ($47). Jefferies issued a "hold" recommendation with a target price of 56,700 won ($42). The average target price from the 12th securities firms is 56,700 won ($42), which is 2.4 times the current stock price.
A key point to watch is the operational schedule for Cape Station, where the first power generation is expected to begin. If electricity production starts as scheduled in the fourth quarter, the company will begin to be re-evaluated from a "loss-making startup" to a "power generator." Until then, Global Investment Information Company Wonrisachi advises that investors need to coordinate their buying timing with the supply-demand pressure (release of locked shares in November) in mind.
in STOCK
