Doosan Enerbility Co., Ltd. (KRX:034020) signed a ₩665.9bn contract on September 1. It will build the Hadong combined-cycle power plant for Korea Southern Power. Construction starts January 2027 and commercial operation is due December 2029. The filing notes there is an advance payment, which the Oman contract covered here two weeks ago did not have.
It is the latest brick on a pile that has grown faster than anything else in the company. The order backlog was ₩26.35tn at the end of June. It was ₩23.05tn at the end of December. Earlier pieces here looked at the equity that rescued the company. They also looked at an offshore wind bid that isn't an order yet. And they covered the cash terms of the Oman deal. This one is about the arithmetic that connects the backlog to the income statement. It examines how fast the book becomes revenue and what that pace costs.
The energy division holds the backlog. It builds reactors, turbines and power plants. Korean press citing the results puts that division's first-half revenue at ₩4.13tn, up 7.4% on a year earlier. The full-year FY2025 figure was ₩7.79tn, about 46% of the consolidated total. The rest of the group is Doosan Bobcat, the US compact-equipment maker, and a construction subsidiary.
Divide ₩26.35tn of backlog by ₩4.13tn of half-year revenue and the book covers 3.2 years of work at the current pace. That is long for a power-equipment company. Siemens Energy and GE Vernova run nearer two years. A long book is good if the margins in it are good. It is a problem if the company can't build fast enough to collect.
The first half's arithmetic checks out. Start at ₩23.05tn, add ₩7.12tn of new orders, subtract ₩4.13tn of revenue recognised, and you get ₩26.04tn. The reported ₩26.35tn is close enough that currency and scope changes explain the rest. Orders in the half were 1.7 times revenue. The book is growing because the company is winning faster than it is building.
The two largest items are nuclear. The Czech Dukovany contract for reactor components and turbine-generators, signed in the fourth quarter of 2025, was about ₩5.6tn. The domestic Shin-Hanul units 3 and 4, signed in 2023, were about ₩3.6tn. Together that is roughly ₩9tn, a third of the book, in two projects that will take most of a decade to deliver.
Gas turbines are the faster-turning part. The company sold five large turbines into the US market in the second half of 2025. It sold seven more in the first half of 2026. Data-centre power demand drove those sales. Middle East combined-cycle projects, including the Saudi Jafurah cogeneration plant and the Oman Duqm plant, sit alongside them. Then there are service contracts and the new Hadong plant.
Nuclear revenue recognises slowly. A reactor vessel forged in Changwon takes years from order to delivery. The revenue comes in as the work progresses. Gas turbines and plant construction move faster. So the mix of the backlog decides the pace. A book that is a third nuclear will convert more slowly than the company's history suggests.
Here is the uncomfortable number. The energy division earned ₩154.4bn of operating income on ₩4.13tn of first-half revenue. That is a 3.7% margin. Korean press reports it as up 69.7% year on year, which is true and which tells you how thin the base was.
Consolidated operating income was ₩547.8bn in the half, ₩233.5bn in March and ₩314.3bn in June. The June margin was 6.7%. Bobcat and the rest of the group earn more than the energy division does.
The company's own numbers on the Hadong contract show the structure. ₩665.9bn is 3.9% of FY2025 revenue. The company will recognise it over three years. Payment comes against progress and delivery milestones. Nothing in the filing says what margin the contract carries, and Korean utility contracts have historically been won on price.
The market is not paying for 3.7% margins. It is paying for what the book might earn once the Czech and domestic reactor work starts to convert at the margins nuclear components have historically carried. That is a reasonable bet and an unproven one.
Building ahead of billing consumes cash, and the statements show it. Consolidated operating cash flow was negative ₩565.5bn in the March quarter and negative ₩321.1bn in June. The half was negative ₩886.7bn. A year earlier the first half was negative ₩584.8bn. The FY2025 total was positive ₩751.8bn, and nearly all of it arrived in the fourth quarter.
That pattern matters. Korean utility and export contracts often settle at year end. If the fourth quarter of 2026 repeats FY2025, the year balances. If it doesn't, the company will have spent another year building a bigger book with borrowed money.
Cash and equivalents fell to ₩2,541.9bn at June 30 from ₩3,090.5bn in March. Current liabilities rose to ₩12.07tn from ₩10.11tn at year end. Korean financial press puts short-term borrowings above ₩4tn. Total liabilities were ₩17.28tn against ₩12.67tn of equity.
None of this is a crisis. The company survived a real one in 2020 and is far better capitalised now. But it means the balance sheet funds the backlog rather than customer advances. Every quarter of order growth adds to the funding need before it adds to the cash.
The company's market capitalisation was ₩50.7 trillion at Friday's close of ₩79,200. Book equity was ₩12.67tn, so the shares trade at four times book. Trailing four-quarter net income is ₩315.1bn, so the earnings multiple is above 160. The backlog of ₩26.35tn is about 52% of the market value.
The shares are 42% below the 52-week high of ₩136,400. That fall happened while the backlog rose ₩3.3tn. The market has, in effect, cut the multiple it puts on each won of backlog. It fell from about 4.7 times to about 1.9 times over the past year. Some of that is the small modular reactor excitement leaving the price. Some of it is the cash flow arithmetic above catching up.
The bear case here could be too impatient. Nuclear components carry higher margins than combined-cycle construction, and the Czech and Shin-Hanul work is only beginning to flow through. If the energy division's margin moves from 3.7% toward 8% as the mix shifts, the same backlog is worth twice as much.
The cash timing could be structural rather than a warning. If fourth quarters always collect what the first three build, the negative half-year figures are noise. Two years of the pattern is not enough to be sure either way.
And the order flow could accelerate again. Management has guided to ₩13.3tn of orders this year and had ₩7.12tn by June before the Oman and Hadong contracts. A US small modular reactor award or a second Czech phase would change the book's size and its nuclear share at once.
The third-quarter results in late October will show whether energy division revenue is rising toward ₩2.3tn a quarter. That would mean the book is converting faster. Or revenue may be holding near ₩2.0tn, which would mean it is not. The backlog figure in the same release will show whether the Oman and Hadong orders pushed it past ₩28tn.
The full-year cash flow statement in February is the decisive one. FY2025 showed ₩751.8bn of operating cash, all of it in the fourth quarter. A repeat means the model works. A shortfall means the company is building the world's longest power-equipment backlog on its own credit.
kstock reads DART every morning and writes up what moved — the contract, the buyback, the number that does not add up. The daily post and a Saturday roundup, by email.