At the end of FY2015, Samsung E&A Co., Ltd. (KRX:028050) — then called Samsung Engineering — reported total equity of negative ₩323,261,968,821.
Not low. Negative. The company owed more than it owned.
That year it recorded revenue of ₩6,441.3bn and gross profit of negative ₩1,089,052,029,479. The work it delivered cost ₩1.1tn more than customers paid for it. Operating loss was ₩1,454.3bn and the net loss ₩1,304.3bn, against a company whose entire equity going into the year was around ₩1tn.
Retained earnings ended FY2015 at negative ₩302.3bn.
The mechanism is visible in one line. Share capital was ₩200,000,000,000 at the end of FY2015 and ₩980,000,000,000 at the end of FY2016 — an increase of ₩780bn in a single year.
That is a rights issue, and it is what put equity back above zero: total equity went from negative ₩323.3bn to positive ₩986.7bn across FY2016. Existing shareholders were diluted enormously; the alternative was insolvency.
Retained earnings then took two more years to cross zero, reaching positive ₩141.0bn at the end of FY2017. From there the climb was steady: ₩688.9bn in FY2020, ₩1,676.5bn in FY2022, ₩3,187.3bn in FY2024, and ₩3,860,050,252,217 at 30 June 2026.
Total equity now stands at ₩4,720.8bn.
Eleven years from negative ₩323bn to positive ₩4.72tn, without a single loss-making year after FY2017.
Nothing signals the end of a recovery like paying shareholders again.
Dividends paid were zero in FY2023 and ₩195,989,352 — under ₩200m, effectively nothing — in FY2024. In FY2025 the company paid ₩129,360,000,000, and the first half of 2026 already shows ₩169,270,416,076 on a cumulative basis.
Behind that is an actual policy rather than a one-off. The company set its FY2024 dividend at 17% of controlling-interest net income and raised the FY2025 figure to 25%, citing operating profit that exceeded its own guidance.
A payout ratio that rises because results beat guidance is the right kind of policy: it shares the upside without committing to a level the company might not be able to hold. It is also a real change of posture at a business that spent a decade retaining every won to rebuild its balance sheet.
The current numbers justify it. Second-quarter 2026 revenue was ₩2,609.3bn, up 19.8% year on year, with operating income of ₩273.1bn, up 51.0%, and net income of ₩182.5bn. Operating margin reached 10.5%.
That follows three years of shrinking revenue — ₩10,624.9bn in FY2023, ₩9,966.6bn in FY2024, ₩9,028.8bn in FY2025 — so the second quarter marks a genuine turn. First-half revenue of ₩4,876.7bn was up 14.0% and first-half operating income of ₩461.3bn was up 36.4%, lifting the half-year margin to 9.5% from 7.9%.
Order intake supports it. The company booked ₩7.6tn in the first half against a ₩12tn full-year target, so more than 60% of the year's target was secured in six months.
The mix is the interesting part: ₩3.4tn in chemical plants, ₩2.4tn in advanced industry — the semiconductor fabs, display plants and biologics facilities it builds for other Samsung affiliates — and ₩1.8tn in what it calls New Energy. Chemical work is driven by Middle Eastern producers expanding investment. New Energy is driven, on the company's own account, by the growth of the artificial intelligence industry and the clean energy transition.
That last connection is worth pausing on. A Korean plant contractor is now describing AI as a demand driver, because data centres need power, and power at that scale needs plants built.
The reason to keep the collapse in view is that the cause has not gone away.
What destroyed the company was fixed-price lump-sum turnkey contracts for Middle Eastern refineries and petrochemical complexes, priced when competition for work was fierce, and then delivered into cost inflation and scope disputes. Under percentage-of-completion accounting, the entire expected loss is booked the moment it becomes probable, which is why a decade of accumulated profit vanished in four quarters.
Samsung E&A is once again winning ₩3.4tn of chemical plant work in the Middle East in a half year. The market is different — oil producers are investing heavily and contractors have more pricing power than in 2013 — but the contract structure that broke the company is the same contract structure it signs today.
The mitigations are real. The advanced industry segment, at ₩2.4tn of first-half orders, is work for affiliated Samsung companies with far lower estimating risk. And the balance sheet can now absorb a bad project: ₩4.72tn of equity against ₩5.78tn of liabilities, of which ₩5.40tn is current, which for a contractor is mostly customer advances and payables rather than debt.
The strongest evidence is eight consecutive profitable years with rising margins. Operating margin went from 0.4% in FY2017 to 7.1% in FY2020, 9.3% in FY2023 and 9.5% in the first half of 2026. A contractor that had systematically underpriced work would not produce that sequence — mispriced contracts surface as margin decay long before they surface as a loss.
Governance has changed too. The company files order announcements continuously — at least eight single-contract disclosures and amendments in the three months to August — which is the kind of transparency that makes a repeat of 2015 harder to hide.
And the guidance-linked dividend policy creates its own discipline. A management team that raises the payout when it beats guidance has an incentive not to buy revenue with bad bids.
Gross margin on the chemical segment as the ₩3.4tn of first-half orders converts to revenue over the next three years. It ran 14.8% at the group level in the first half of 2026. Chemical plant EPC done well earns high single digits; done badly it earns nothing at all, and the difference does not appear until year two.
Second, the FY2026 payout. The policy moved from 17% to 25%. Whether it holds at 25% or ratchets further tells you how confident the board is in the order book it has just signed.
Third, the fourth quarter. Korean contractors review contract costs annually and take the adjustment then. Samsung E&A has now posted several clean fourth quarters — ₩277.4bn of operating income in the fourth quarter of 2025 and ₩295.8bn in the fourth quarter of 2024. A fourth quarter that breaks that pattern would be the first real test of whether the estimating discipline holds.
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.