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.
Today, no essay. A scoreboard. This is not financial advice — a view, not a recommendation to buy or sell anything.
Two numbers that don’t fit in the same sentence
Type “south korea gas turbine market” into Google and you get something around four hundred million dollars. MarkNtel Advisors sizes South Korea’s gas turbine market at about $432 million in 2025, growing 5.08% a year to $611 million by 2032. Grand View Research sizes the same market at $229.7 million (2024), growing at 2% to $275.9 million by 2033. Both numbers say the same thing: South Korea’s gas turbine market is small and boring.
In March, Doosan Enerbility (Korea Exchange: 034020 — think of it as Korea’s heavy-power-equipment champion) signed a contract with a US big-tech firm for seven 380MW-class gas turbines, worth roughly ₩1 trillion (~$665 million). One contract is bigger than “the entire South Korean gas turbine market.”
Both numbers are correct. The rulers are different. The market researchers measure how much Korea buys. Doosan’s contract measures how much Korea builds. Right now the money is moving on the second ruler.
Scoreboard — two rulers
| What | Number | Ruler |
|---|---|---|
| South Korea gas turbine market (2025e) | $432M → $611M (2032), 5.08%/yr | What Korea buys |
| Same market, other estimate (2024) | $229.7M → $275.9M (2033), 2%/yr | What Korea buys |
| Doosan single contract (Mar 2026, 7× 380MW-class, US big tech) | ~₩1tn (~$665M) — about 2.7GW of capacity | What Korea builds |
| Doosan Q1 2026 gas turbine orders | 10 units (3 domestic / 7 for North American data centers) | Builds |
| Doosan cumulative US gas turbine contracts | 12 units | Builds |
| Doosan order backlog (end-Mar 2026) | ~₩24tn (~$16B) — up from ₩13tn (~$8.6B) at end-2022 | Builds |
FX basis ₩1,503.4/$ (as of 2026-07-14, consistent with our prior pieces). Verify figures against Doosan IR and KRX at the time of reading.
The twist is the order of magnitude. “A $432M market” and “a ₩24 trillion (~$16B) backlog” describe the same country, the same industry, the same company — and they are roughly 37× apart. (Fair caveat: the backlog is Doosan’s whole book, not gas turbines alone.) That search takes you to a number showing a low single-digit percentage of the story. The rest lives in an order book the domestic-market number was never built to see.
🎩 Under the Gat — Grand View literally counts Korea as 2.1% of the global gas turbine market — a slice of what the world buys. Neither report ever claimed to measure what Korean factories ship to Texas. (Honest flip side: that $432M isn’t Doosan’s either — Korea’s own plants buy from GE, Mitsubishi, and Siemens too. The home market is small and shared.) Put the wrong words in the search box and you don’t get a wrong answer — you get a correct answer to a question nobody asked. You showed up at a factory holding a spreadsheet that counted Korea as a customer.
Why this is happening now — the American scoreboard
Here’s the part that matters even if you never touch a Korean stock.
| What | Number |
|---|---|
| GE Vernova turbines on order (end-2025) | 83GW — up from 62GW one quarter earlier |
| Its factory output, 2028 target | 24GW/year → that backlog is ~3.5 years of production, already spoken for |
| Slots sold out through | 2029; 2030–31 under negotiation |
| Data center / AI share of backlog | ~20% — and 9GW of it is slot reservations: customers paying up front, at prices above the existing backlog, just to hold a place in line |
Verify against GE Vernova IR.
Read that fourth row again. Customers are paying a premium for the right to wait. That is what a genuine shortage looks like — not a price move, a queue.
So “sold out until 2029” isn’t a sentence about GE Vernova. It’s a sentence about everyone who needs electricity in 2029. And it raises the only question that matters: where do the orders bumped from that queue go?
One line in Doosan’s Q1 answers it: of 10 gas turbines booked, 7 went to North American data centers. For scale, that single March contract — 2.7GW — is roughly six weeks of GE Vernova’s entire targeted 2028 output, placed with a company most US investors have never heard of.
The old line says the pick-and-shovel seller gets rich in a gold rush. That line is stale. The real question is: when the pick shop is sold out, who’s the shop next door?
Today’s tape (July 16, 2026 close)
| What | Number |
|---|---|
| KOSPI (Korea’s main index) | 6,820.60 (−6.37%) — after +6.24% the day before. A round trip in two sessions, with a sidecar (an automatic brief halt when sell orders cascade) triggered |
| Sectors | Chip-heavy electrical & electronics −9.43% vs machinery & equipment (where Doosan sits) −4.70% — chip names took roughly twice the hit |
| Doosan Enerbility | ₩73,000 (~$48.6) → ₩69,700 close (−4.52%, ~$46.4) |
On a day when “the AI bubble” story hit semiconductors, the power-equipment side of the same AI theme took only half the hit. This proves nothing — it’s one day’s number. But it’s a hint that the market doesn’t sell “the AI trade” as a single block.
So can you actually buy it?
Now the deflating part, and you deserve it straight: not in your brokerage app. Doosan has no US-listed ADR. The routes are a broker offering direct KRX (Korea Exchange) access, or a Korea ETF like EWY, where the weighting is a rounding error. Which means the only name in this piece you can trade on a US exchange is the one that’s sold out.
That friction is exactly why this story stays invisible from here — and why it’s worth knowing anyway. The queue behind GE Vernova’s 2029 wall is where the AI power trade actually gets decided, and a chunk of it is being built in Korean factories whether or not there’s a ticker for it in your app.
🎩 Under the Gat — the other side. Doosan is a ~₩44.7tn (~$29.7B) company trading near 290× earnings (verify basis at KRX) — that’s what a thin earnings base does to a multiple, and it only makes sense if you’re underwriting 2029 rather than 2026. It means “the earnings haven’t arrived yet,” and it also means “too much has already arrived,” and the scoreboard can’t tell you which. First deliveries on that March contract start around May 2029, one unit a month. A backlog is revenue visibility, not a margin guarantee — material costs, FX, and tariffs all move after the signature. We flagged this same trap before with the transformer names in the AI power valuation gap. A scoreboard doesn’t answer “cheap or expensive.” The 2029 delivery schedule does.
🎩 And one caveat. Doosan is not “Korea’s GE Vernova.” GE Vernova carries gas, nuclear services, and wind, at a different weight class entirely. Doosan is concentrated in power equipment and nuclear fabrication — including small modular reactors (SMRs). It’s passing through the same bottleneck from a different angle — the shop next door, not the same shop.
The one-line takeaway
Don’t search “south korea gas turbine market.” Search “who builds them when GE is sold out.” The answer to the first question is $432 million. The answer to the second is sitting inside a ₩24 trillion (~$16B) order book. (Again: not financial advice. Every figure here should be re-verified against Doosan IR, KRX, and GE Vernova IR before you act on anything.)
This is not financial advice. Won-to-dollar conversions use ₩1,503.4/$ (2026-07-14 basis). Preliminary and estimated figures — backlog, market cap, P/E — should be re-verified at the time of reading against primary sources (Doosan Enerbility IR, DART — Korea’s electronic disclosure system — KRX, and GE Vernova IR).
If you want the rest of the queue mapped — who supplies the buildout, layer by layer: Korea’s power generation stocks — gas turbines, SMR · The AI data center power supply chain · The AI power valuation gap.
Frequently Asked Questions
Why are gas turbines sold out until 2029 — and who else can build them?
AI data center power demand is growing faster than new generation can be built. GE Vernova reported roughly 83GW of gas turbine backlog at end-2025, with production slots effectively booked through 2029. Only a handful of firms design and build large-frame turbines; South Korea’s Doosan Enerbility is one, and booked 7 units for North American data centers in Q1 2026 alone. Verify current figures. This is not financial advice.
Is South Korea’s gas turbine market really only about $400 million?
That number measures the South Korean domestic market — turbines sold into Korea. Grand View Research sizes it at $229.7M (2024) and explicitly counts Korea as ~2.1% of the global gas turbine market; MarkNtel puts it at ~$432M (2025), with Korean power plants as the end users. Korean makers’ exports are not counted there. Doosan’s single March US contract (~₩1tn / ~$665M) exceeds either domestic estimate.
Does that mean the market research is wrong?
No — it answers a different question. The reports measure what Korea buys and installs; they do not measure what Korean factories sell abroad. If you want the export story, the domestic market-size number is simply the wrong ruler.
Can US investors buy Doosan Enerbility?
There is no US-listed ADR. The practical routes are direct KRX access through a broker that offers it, or indirect exposure via a Korea ETF such as EWY — though the weighting there is small. Verify holdings and access before acting.
Is a ₩24 trillion (~$16B) backlog automatically bullish?
No. A backlog is revenue visibility, not a margin guarantee. Deliveries on the March contract are expected to begin around May 2029, and material costs, FX, and tariffs move after a contract is signed. Whether contracts are fixed-price or indexed shows up in quarterly results. Not financial advice.
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.