The gas turbine industry, currently controlled by an oligopoly of three companies (Japan’s MHI, Germany’s Siemens, and America’s GE Vernova), is experiencing a new golden age due to a significant increase in demand generated by the AI industry’s hunger for electricity. And to think that just a few years ago, the gas turbine sector was considered dead.
US utilities and artificial intelligence developers have launched a coordinated rush for gas turbines, driven by OpenAI and SoftBank’s $500 billion Stargate project to invest in artificial intelligence infrastructure, including gas-fired power plants.
The US is expected to account for 46% of global gas turbine orders this year, up from a recent historical average of 29%. Data center developers are even developing grid-independent gas-fired power plants. Mitsubishi Power has doubled its orders, and for other manufacturers, the increase in orders appears to be even greater, so much so that manufacturers are now requiring deposits to reserve spots in the queue (a sales scheme already used during another boom in the early 2000s) in addition to applying surcharges for early deliveries. In April, GE Vernova said it had largely sold out its inventory for 2026 and 2027. It is now negotiating with customers for delivery in 2030.
Mitsubishi Power expects high demand for at least five to ten years, due to strong demand for artificial intelligence. It believes this phase is fundamentally different from the gas turbine market crisis around 2000, which was caused by the deregulation of the US energy sector.
As energy producers seek to increasingly meet the demand for artificial intelligence in the US, there is a real risk of neglecting the next driver of natural gas consumption growth: emerging markets. These countries need energy to fuel their development, which depends only minimally on investment in AI. In fact, at this moment, the demand for gas turbines to power AI development poses a serious supply problem for developing countries. In the short term, this will lead them to favor fossil fuels, where China certainly dominates in Southeast Asia. All gas turbines have been purchased by Americans, Europeans, and Middle Eastern countries; it will be very difficult for emerging Asian economies to increase their exposure to natural gas.
This is even more true for non-fossil fuels, where nuclear power plays a major role. The focus of those involved in AI projects, at least the most prominent ones, is on micro-reactors, some of whose technical characteristics we outlined in our August 8, 2025, in-depth analysis, in which we also discussed the main players in this field. Today, history and performance lead us to talk about one of these players: Oklo Inc.
Oklo—named after the location in Gabon where the only natural example of nuclear fission occurred—is a company with no revenue, no licenses to operate reactors, and no binding contracts for energy supply.
But that hasn’t stopped the Silicon Valley-based startup from riding the wave of investor enthusiasm that has pushed its stock market valuation to over $20 billion, an increase of more than 500% since the beginning of the year. The company, backed by Sam Altman (who was its president and oversaw its IPO) and with close ties to Donald Trump’s energy secretary, has set ambitious goals to supply commercial power to its first customers in 2027, after launching its pilot project in Idaho last month.
Oklo, led by husband and wife team Jacob and Caroline DeWitte (both PhDs in nuclear engineering from MIT), envisions a future powered by a new generation of small modular reactors that use liquid sodium instead of water as a coolant. The company aims to become a leader among companies that will provide energy-intensive data centers with the power they need to fuel the AI boom.
However, the surge in its stock, fueled by enthusiasm from retail investors who represent a disproportionate share of its shareholders, has worried experts, who fear the stock is overheating. It is among the highest-valued pre-revenue companies listed in the United States.
Oklo has recently experienced significant price retracements (well into double digits), reflecting concerns about its exorbitant valuation.
Oklo’s stunning performance has fueled criticism from those who claim that it has benefited from ties to the US administration, which has supported it for several federal programs. The Secretary of Energy is a former member of Oklo’s board of directors.
In May, Trump invited DeWitte to the White House for an event in which the president pledged to quadruple the United States’ nuclear capacity by 2050, commissioning 10 new large reactors from Westinghouse by executive order (see our Insight of June 27, 2025). Although Westinghouse is clearly the main beneficiary of the White House’s attention, observers have noted that the Department of Energy has selected Oklo for programs aimed at accelerating the construction of low-carbon reactors (SMRs) and nuclear fuel manufacturing facilities, and has committed to providing it with specialized and scarce reactor fuel, while considering granting it access to weapons-grade plutonium to produce its fuel. This is precisely Oklo’s strength: it produces—or claims to produce—fuel for micro-reactors. This vertical integration, combined with political support, has helped give Oklo an advantage over its rivals and explains why Bank of America analysts value the company higher than fellow SMR developers NuScale Power and Nano Nuclear Energy.
The conflict of interest worries not only the opposition, but also investors and shareholders, 13% of whom sold the stock short, convinced that the DeWittes had underestimated the time and money needed to commercialize the technology. The US Nuclear Regulatory Commission’s decision to reject Oklo’s previous application to build a sodium-cooled reactor in 2022 has also raised questions. Some nuclear experts have highlighted the failure of sodium-cooled reactors built in the US between 1950 and 1976: liquid sodium is highly corrosive, flammable, and explosive when in contact with air and water. Many countries have already attempted to build these reactors but have failed to demonstrate their commercial viability on a large scale.
But rather than conflicts of interest, politicians should be more concerned about the risks of proliferation, as Oklo’s plans would see plutonium pass into the hands of private industry, where it could be at risk of diversion or theft by those seeking to build an atomic bomb.
The DeWittes insist that sodium cooling technology has made great strides and offers a safety advantage over other types of reactors because it does not require such high pressure. This can reduce costs, a factor that has previously hampered nuclear projects. After all, TerraPower, the nuclear company founded by Bill Gates, and Aalo Atomics are pursuing similar technology.
Altman left the board in April 2024 to avoid any potential conflicts of interest, while Oklo stepped up talks with artificial intelligence companies about energy contracts. Oklo has so far signed non-binding memoranda of understanding with Big Tech customers, but no legally binding power purchase agreements, a further source of concern for some analysts.
And then there are the shareholders. Not only those who sell short, but also the army of small shareholders who certainly do not lend solidity to an innovative project that requires solidity; the typical solidity provided by institutional investors or traditional nuclear financing agreements.
We greatly appreciate innovative projects, provided that there is no mixing with politics and that intangible assets (patents, solid supply contracts, etc.) are present, if not actually included in the balance sheet. Another point is the devaluation of the use of micro-reactors, which, judging by the history of Oklo, seem destined to power the artificial intelligence industry, when in fact they have many other uses where micro-reactors are an optimal solution (such as supplying remote regions).
Disclaimer
This post expresses the personal opinion of the Custodia Wealth Management staff who wrote it. It is not investment advice or personalized advice and should not be considered an invitation to carry out transactions on financial instruments.
