This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.
Everyone is buying the AI chip. Fewer people ask the next question: once the data center is built, who actually makes the electricity? Increasingly, the hyperscalers are buying their own power plants — and the machines that make the power, large gas turbines, are sold out for years. That shortage is how a Korean heavy-industry company became an unlikely AI-power play. This is not financial advice.
📍 Part of TheGatBull’s AI Power Supply Chain series — the full map of where Korea sits in the AI electricity build-out. This piece zooms into power generation — gas turbines & SMR.
This is Episode 2 of our series on the AI data center power supply chain — a close-up on generation: gas turbines and small modular reactors.
The short answer: Doosan is Korea’s purest “AI power crunch” bet — with a catch
Doosan Enerbility (KRX: 034020) makes both ends of power generation: the gas turbines the world needs now, and the small modular reactors (SMRs) it will want in the 2030s. The whole thesis rides on timing. The global gas-turbine “Big 3” — GE Vernova, Siemens Energy and Mitsubishi — are effectively sold out, with order backlogs running about five years and some delivery slots already booked into the 2030s. Doosan is, in effect, the world’s only meaningful alternative to that trio for large gas turbines — so it catches the overflow. The catch: buying the stock today means buying deliveries that land around 2029 and SMRs that arrive in the 2030s, and the market has already paid forward much of that future.
Why now — AI wants its electricity immediately
AI data centers eat power, and they eat it fast. The problem is that generation equipment can’t keep up. Only three companies can really build large gas turbines at scale — GE Vernova, Siemens Energy and Mitsubishi — and together they hold well over 75% of the market, sold out for years. For a hyperscaler that has finished a data center but has no electricity to run it, that’s a nightmare, which is why they’re increasingly buying turbines themselves and building power plants right next to the racks. And when the Big 3 are sold out, the only remaining option is a fourth supplier. That’s where Doosan walks in: in 2019 it became the first Korean firm to validate its own large gas turbine, making it one of the very few companies outside the Western trio that builds them with its own technology.
🎩 Under the Gat — Wall Street calls this a “turbine shortage.” Korea just calls it ppalli-ppalli (빨리빨리, the country’s “hurry-hurry” speed culture). When the Big 3 are sold out, what Doosan really sells isn’t the best turbine — it’s the one you can actually get. That’s the point outsiders miss.
The bull case — the “overflow” became a record order book
The numbers back the bulls. Doosan reportedly booked roughly ₩14.7 trillion (~$10 billion-plus) of new orders in 2025 — more than double the prior year — and guided to about ₩16.3 trillion (~$12 billion) for 2026, lifting its gas-turbine assumption from 11 to 16 units. The symbolic deal: a seven-unit, 380 MW-class gas-turbine contract with a US big-tech buyer, the company’s largest single order ever, with deliveries of about one unit a month starting May 2029. Industry watchers peg the buyer as Elon Musk’s xAI, though Doosan hasn’t confirmed it.
Then there’s the second engine: SMRs. Doosan has a binding reservation for the steel components of 16 of X-energy’s Xe-100 reactors and is building a dedicated SMR plant in Changwon (reportedly around ₩80 billion, ~$55 million) aimed at roughly 20 units a year by 2031. If gas is “the power for now,” SMRs are “the clean baseload for the 2030s.”
🎩 Under the Gat — Doosan’s real appeal is that it runs on two clocks. Gas turbines are the cash flow of the next three to five years; SMRs are the option on the decade after. One stock, a near-term catalyst and a long-term story bolted together. Tempting picture — the question is the price.
The bear case — a latecomer, and a number called “2029”
Now the other side, fairly. The skeptic case rests on three points. First, Doosan is a latecomer. Building its own large gas turbine is impressive, but it only entered the field in 2019, the world’s fifth player. GE, Siemens and Mitsubishi carry decades of fleet hours and a track record on efficiency and reliability. If the main reason a hyperscaler picked Doosan was “the Big 3 are sold out” rather than performance, that edge softens the moment the Big 3 add capacity. Second, orders aren’t revenue. There’s a long lag between a glamorous backlog and actual profit — gas-turbine deliveries start around 2029, and nuclear and SMR work can take five to ten years from order to recognized revenue. Third, it’s pre-priced. The stock has already re-rated hard and trades near multi-year highs, and because this is fundamentally a cyclical business, a small disappointment can trigger an outsized pullback.
🎩 Under the Gat — A backlog is a promise, not cash. Three things actually matter here: the margin when gas-turbine revenue starts landing around 2029, whether Doosan keeps winning orders once the Big 3 expand, and the moment SMRs turn from an option into earnings. Watch those, not the order headlines.
The anchor for global readers: Korea’s GE Vernova, a generation behind
| Doosan Enerbility (Korea) | GE Vernova (US anchor) | |
|---|---|---|
| Business | Gas turbines + nuclear/SMR main equipment + power EPC | Gas, wind, electrification — a pure-play power giant |
| Gas-turbine position | Entered 2019, the world’s 5th challenger | One of the Big 3, decades of fleet hours |
| Moat | Almost the only large-turbine alternative outside the West, plus a Korean nuclear hub | Track record, share, huge data-center backlog |
| Decisive difference | “Available now” capacity is the weapon | “Proven performance and volume” is the weapon |
Not a perfect parallel — GE Vernova is a pure power player with decades of operating data; Doosan is a diversified heavy-industry group that only entered large gas turbines in 2019. Buying Doosan isn’t buying the No. 1 — it’s betting on the second source in a sold-out market.
The satellites: KEPCO E&C and KEPCO KPS
Widen the lens and two KEPCO affiliates join the theme. KEPCO Engineering & Construction (KRX: 052690) does reactor systems design; KEPCO KPS (KRX: 051600) handles commissioning and maintenance. Both benefit as nuclear orders — large reactors and SMRs — rise, but they lag Doosan in the cycle, since design and service revenue lands at the construction and operating stages. The near-term momentum sits with Doosan; the longer nuclear cycle is where the KEPCO names spread the exposure.
The policy backdrop: Korea’s pragmatic nuclear turn
The fate of Doosan’s SMRs and the KEPCO names ultimately rests on policy — and in Korea, nuclear has historically flipped 180 degrees with each administration (Moon Jae-in’s phaseout, then Yoon Suk-yeol’s nuclear push), which is exactly why foreign investors long treated it as a symbol of “policy risk.” The Lee Jae-myung government has chosen a third path: growing nuclear and renewables at the same time, a pragmatic line.
Recent developments in 2026:
- Initially cautious, President Lee shifted after public surveys (over 80% see nuclear as necessary) and surging AI- and chip-driven power demand — effectively ending the phaseout stance.
- Korea confirmed new reactor sites for the first time in 14 years: two large reactors (1,400 MW each, targeted for 2037 and 2038) in Yeongdeok, and the country’s first innovative SMR (i-SMR) candidate site in Gijang — following the 2024 restart of the Shin-Hanul 3 and 4 units.
- An SMR Special Act passed in February 2026, setting up a five-year plan and a promotion committee, with the government reportedly committing about ₩1.2 trillion (~$830 million) by 2030 to domestic SMR designs.
- Energy-mix targets put nuclear at roughly 32% of generation by 2030, rising toward 36% by 2038, with renewables at 22% then 33%.
For investors, that hands Doosan (reactor equipment and SMRs) and the KEPCO names (design and maintenance) a domestic-demand tailwind backed by policy. More importantly, it signals policy continuity — a government that was once cautious turning pragmatic chips away at the old “nuclear dies when the government changes” discount. The caveat, fairly stated: environmental groups have pushed back, and unresolved issues like a radioactive-waste repository and grid reinforcement remain — and “continuity” is still on trial.
🎩 Under the Gat — The real risk in Korean nuclear stocks was never the technology — it was the politics, a switch each administration flipped on and off. The point of this pragmatic turn is that it makes the switch wobble less. Less, not never. The true signal is when someone finally breaks ground on the waste site.
The bottom line: in a world where gas turbines are sold out and AI wants power yesterday, Doosan is Korea’s most direct generation play — running on two clocks, gas now and SMRs later, with a friendlier policy backdrop than Korean nuclear has had in years. Just remember what you’re buying: deliveries that start around 2029 and a valuation that has already reached for that future. The opportunity is real; when and at what price matter as much as which name. A view, not advice.
— Mr. Gat 🐂
This is not financial advice. Past performance does not guarantee future results, and all figures should be checked against primary sources as of your trade date.
Frequently Asked Questions
Is Doosan Enerbility a nuclear stock or a gas-turbine stock?
Both. Its near-term catalyst is gas turbines, with deliveries on its record US order starting around 2029, while its long-term story is nuclear and small modular reactors (SMRs) in the 2030s. Holding two different time horizons in one stock is both its appeal and its complexity. This is not financial advice.
Is the reported xAI contract confirmed?
Doosan disclosed a record seven-unit gas-turbine order, but the idea that the buyer is Elon Musk’s xAI is an industry estimate, not a company confirmation. Treat the counterparty as unconfirmed.
How can foreign investors access these names?
Doosan Enerbility (KRX: 034020), KEPCO Engineering & Construction (KRX: 052690) and KEPCO KPS (KRX: 051600) all trade directly on the Korea Exchange through brokers that offer Korean market access, such as Interactive Brokers. Account and tax details are covered in our investing-in-Korea guides.
What is the biggest risk?
Timing and pre-pricing. The order book is dazzling, but revenue recognition is years away — gas-turbine deliveries begin around 2029 and nuclear runs longer. The shares have already re-rated, so delivery delays, thin margins or rate moves can sting. Verify all figures against primary sources as of your trade date.
Does Korea’s nuclear policy actually support this?
Increasingly, yes. The Lee Jae-myung government has shifted to a pragmatic stance — confirming new reactor sites, passing an SMR Special Act, and targeting nuclear at 32% of the power mix by 2030 rising toward 36% by 2038. That reduces the historic risk that Korean nuclear policy flips with each administration, though waste-site and grid challenges remain.
This article is for informational purposes only and is not financial advice. TheGatBull may earn a commission from some links at no cost to you — see our disclosure and full disclaimer.