Doosan Enerbility Co., Ltd. (KRX:034020) is a survival story. In 2020, as Doosan Heavy Industries, it needed a government-arranged credit line to stay solvent, sold assets, shrank, and renamed itself. Today it is one of a handful of companies on earth that can forge a complete nuclear steam supply system, and it sits in the middle of two demand stories that are as good as heavy engineering gets: nuclear construction restarting after two decades, and data centres needing generating capacity faster than anyone can build it.
The market has noticed. At the August 27 close of ₩85,900 across 640,561,146 shares, the company is capitalised at ₩55.02tn.
Book equity at June 30 was ₩12.67tn. That is 4.3 times book, for a heavy machinery manufacturer.
Before deciding whether that is reasonable, it is worth looking at who paid for the recovery that produced it.
The clearest record of the rescue is the share capital line, because it only moves when shares are issued.
FY2015 and FY2016: ₩596,808,980,000. FY2017: ₩596,836,515,000. FY2018: ₩650,255,065,000. FY2019: ₩1,075,255,425,000. FY2020: ₩1,937,707,325,000. FY2021: ₩2,675,624,980,000. FY2022: ₩3,256,061,215,000. FY2023 through today: ₩3,267,327,000,000.
From ₩596.8bn to ₩3,267.3bn in eight years. Five and a half times. Each of those steps is an issuance, and every issuance dilutes whoever held before it.
That is what saving this company cost. Not a restructuring of debt, not a cleanly executed turnaround funded by operations. New equity, repeatedly, at prices that reflected the distress of the moment.
The usual way this story gets told is that Doosan Enerbility deleveraged. It did, in ratio terms, and the ratio flatters the mechanism.
Total liabilities were ₩20.23tn at the end of FY2015 and ₩17.28tn at June 30, 2026. That is a decline of 15% over more than a decade. Total equity went from ₩7.03tn to ₩12.67tn, up 80%.
So the debt-to-equity ratio improved from 2.88 times to 1.36 times, which reads as dramatic. Almost all of it came from the denominator. The company did not pay down much; it raised enough equity to change the shape of the balance sheet around a debt load that is still substantial.
For a lender that distinction is irrelevant. For a shareholder it is the whole story, because it determines who bore the cost.
One more line makes the point. Retained earnings at the end of FY2020 were negative ₩2,046.6bn, an accumulated deficit built up over years of losses: FY2015 lost ₩1,750.9bn, FY2016 ₩215.5bn, FY2017 ₩198.0bn, FY2018 ₩430.1bn, FY2019 ₩223.6bn, FY2020 ₩838.4bn.
At the end of FY2021 retained earnings were positive ₩773.2bn. Net income that year was ₩645.8bn.
A ₩2.82tn swing on ₩646bn of profit. An accumulated deficit does not shrink by more than the profit earned unless it is offset against capital, which is precisely what Korean law permits and what a company emerging from restructuring routinely does. The deficit was cancelled against paid-in capital rather than worked off through operations.
Nothing improper about it. But anyone looking at Doosan Enerbility's retained earnings today, at ₩1,874.1bn, should know that the number restarted from a reset rather than accumulating continuously.
Which brings us to the present, and the awkward part.
Net income was ₩517.5bn in FY2023, ₩394.7bn in FY2024 and ₩205.2bn in FY2025. Down each year, by 24% then by 48%. Operating income over the same three years went ₩1,467.3bn, ₩1,017.6bn, ₩762.7bn, with operating margin falling from 8.34% to 6.27% to 4.47% on revenue that was essentially flat at ₩17.59tn, ₩16.23tn and ₩17.06tn.
Three consecutive years of declining profitability, during the period when the nuclear and power-demand narrative was strengthening.
The first half of 2026 broke the trend. Operating income was ₩547.8bn against ₩413.6bn a year earlier, up 32.4%, with the margin recovering to 6.10%. That is genuine improvement and it is one half of one year.
There is also a composition problem in the earnings. Doosan Enerbility consolidates Doosan Bobcat without owning all of it, so reported net income includes amounts belonging to minority shareholders, and intangible assets of ₩9.10tn exceed property, plant and equipment of ₩5.92tn on the balance sheet, which is goodwill from that consolidation rather than forges and factories. Anyone valuing the nuclear business off consolidated numbers is valuing an American compact equipment maker at the same time.
Annualise the first half's ₩286.6bn of net income and you get about ₩573bn, so roughly 96 times the market capitalisation. On FY2025's actual ₩205.2bn it is 268 times. At 4.3 times book, the shares price a company whose earnings must multiply several times over.
The bull case is not stupid. Nuclear new-build orders are long-dated and enormously valuable, Doosan holds forging capacity that cannot be replicated quickly, and gas turbine demand from data centres is real and urgent. If the order book converts at even historical margins on a larger revenue base, the earnings arrive.
The bear case is that this is the same company that has posted seven loss-making years out of the eleven on file, whose operating margin fell for three consecutive years into a strong demand environment, whose operating cash flow is negative in every interim period, and which just signed a ₩928bn Oman contract with no advance payment.
Operating margin, quarter by quarter, rather than order announcements. Orders are the easy part in this cycle; converting them at a margin that justifies 4.3 times book is the hard part. The first half's 6.10% needs to keep climbing toward the 8.34% of FY2023 and beyond.
The second marker is the share capital line in the FY2026 annual report. It has been unchanged at ₩3,267,327,000,000 since FY2023, which is the longest stretch without an issuance in a decade. If it stays flat through another year of heavy working capital demands and a growing order book, the company is finally funding itself. If it rises again, the pattern that built this balance sheet has not ended, and shareholders will pay for the next expansion the way they paid for the last rescue.
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.