The best surgeon in the country does not decide when she operates; the hospital schedule does. On days with no surgery the equipment lease is still due and the team is still paid. Buyers of such an asset price the skill. The variable that sets the price is utilization, and someone else decides it.
That is the structural question inside GS Engineering & Construction’s reported review of Dong-Ah Geological Engineering, a KOSDAQ-listed Korean specialist contractor ranked first domestically in shield tunnel boring machine (TBM) work and marine deep cement mixing. The seller, Crescendo Equity Partners, is reportedly asking for a multiple on adjusted EBITDA of roughly KRW 38 billion. Reported 2025 operating profit is roughly KRW 18 billion. The KRW 20 billion in between is the deal.
Nothing is signed. A report on 11 September 2026 said the sale was progressing slowly; a report on 15 September said GS E&C was reviewing an acquisition. Four days between contradictory signals is the most accurate available indicator that no agreement exists.
The Deal
- Target: Dong-Ah Geological Engineering (Korea), specialist ground engineering and tunnelling. 2025 revenue approx. KRW 467.7 billion; reported operating profit approx. KRW 18 billion (2024: KRW 9.6 billion; 2019: KRW 22.9 billion). Market capitalization approx. KRW 220 billion on 23 September 2026.
- Seller: Crescendo Equity Partners via the SPC Dover Holdings, approx. 34%, largest shareholder. August 2019 entry: founder shares at KRW 18,000 for approx. KRW 40.3 billion (an 18–19% stake), plus a KRW 20 billion convertible bond and a KRW 20 billion bond with warrants — approx. KRW 80.3 billion in total (about USD 59 million at KRW 1,365/USD, 23 September 2026).
- Prospective buyer: GS E&C, a counterparty since 1987. 2025 revenue approx. KRW 12.45 trillion, operating profit approx. KRW 438 billion, net income attributable to shareholders approx. KRW 94 billion after a large one-off loss at a water-treatment subsidiary.
- Reported ask: adjusted EBITDA of approx. KRW 38 billion at 6–7x plus net cash — cited by some outlets at around KRW 300 billion (approx. USD 220 million), by others at the mid-KRW 300 billions or higher. A first attempt in 2025 was shelved over price; the process reportedly resumed around June 2026.
The divergence is itself a datum. KRW 38 billion at 6x is KRW 228 billion and at 7x KRW 266 billion; reaching the mid-300s needs net cash of roughly KRW 80–90 billion, reaching 300 needs KRW 40–70 billion. Subtracting an assumed KRW 80 billion of net cash from the market capitalization leaves an implied enterprise value of about KRW 140 billion, so the ask carries a 60–90% premium to the market’s mark — 45–65% at KRW 60 billion of net cash. Either way it sits above the customary 20–40% control premium.
Structure Breakdown
Domestic leadership in shield TBM fills the first page of a teaser; it does not set the price. One variable does: whether the number the company claims to earn is cash that remains with shareholders. The KRW 20 billion gap between adjusted EBITDA and reported operating profit has two possible identities. If it is mostly depreciation and amortization, the question only relocates: in an equipment-based business depreciation is the present notation of a future replacement obligation, and if maintenance capex runs near it, real free cash flow falls below KRW 20 billion and a 6–7x headline becomes a double-digit multiple on cash. That is arithmetic on an assumption, unsettled until actual capex and fixed-asset book values are examined.
If instead the gap is largely the adjustment itself — one-off reversals, excluded losses, additions labelled normalization — the problem is worse: the seller has not disclosed the items, and the numerator the multiple is applied to is unverified. The architect doubts the numerator and suspects the denominator. Here both deserve it, since the move from 18–19% to 34% implies the mezzanine was largely converted, and the diluted share count has to be fixed before any per-share figure means anything.
Then the seller’s clock. Seven years have passed since August 2019, with one failed attempt and one restart; for a fund, time is a direct deduction from internal rate of return. The share price complicates the floor without removing it: the 52-week high of KRW 22,500 on 16 September followed the acquisition report, but the stock had already moved from a 9 June low of KRW 13,200 into the 17,000s by early September on first-quarter results, the June restart, and power-infrastructure expectations. In a listed-company takeover the market price is the floor of the negotiation.
On the buyer’s side the weight sits in a shrinking core: 2026 guidance of KRW 11.5 trillion in revenue and KRW 17.8 trillion in new orders is below 2025, housing revenue is in structural decline, and civil revenue is estimated at approx. KRW 1.33 trillion against approx. KRW 1.46 trillion in 2025. The vertical-integration story — bring a 38-year subcontractor in-house and control the high-difficulty work, as on the KRW 1.7 trillion Singapore T301 depot project where the two reviewed methods together from the bid stage — rests on one undisclosed number: the GS-bound share of Dong-Ah’s KRW 467.7 billion revenue. If the share is high, consolidation relocates profit rather than creating it. If it is low, third-party revenue is a clear addition and the deal’s quietest risk, because competing general contractors may decline to award high-difficulty work to a GS E&C affiliate. The more net-new revenue there is, the more of it can leave; a subcontractor’s customers become its parent’s competitors on the day it is consolidated.
The Governing Logic
An architect builds a structure not to earn but to not lose. Read that way, the standoff is not a fight over a number; it is an allocation of who carries future uncertainty, and it begins with the 2019 design. Crescendo deployed KRW 80.3 billion and put only KRW 40.3 billion into shares. The other KRW 40 billion went into mezzanine. Its downside protection does not come from the conversion price, which generally reflects the market at issuance; it comes from the claim. Equity is recovered only if the company does well; a claim is recovered as long as the company does not fail, and it ranks ahead at liquidation. Half the money was kept as a claim, which pre-wrote the scenario in which the share never clears the entry price.
That scenario arrived. At the end of 2023 the stock sat around KRW 14,000, more than 23% below the KRW 18,000 purchase price; it has recovered into the 19,000s, still short of the approx. KRW 21,000 mezzanine issue price. The constraint is precise: common stock at market cannot deliver a meaningful fund return, so a private control sale with a premium is the only real route out. A KRW 300 billion-range ask reads less as greed than as the floor beneath which the fund’s return does not exist. The seller is structurally limited in how far it can cut; the buyer is structurally short of grounds to pay. Two failures in two years are what non-overlapping constraint sets look like, and only one thing clears them: the single variable called price is broken into several conditional ones. That is the operating point of Proof-Deferred Pricing — the seller’s number is granted, contingent on the seller’s number arriving.
Downside Protection
The reviewing party is a strategic buyer, which changes the test from IRR to after-tax return on invested capital. KRW 18 billion of operating profit at an assumed 24% tax rate is approx. KRW 13.7 billion: high-3% against a mid-KRW 300 billion payment, around 5% against approx. KRW 260 billion net of acquired cash, and short of a construction cost of capital assumed in the 9% range — with synergy set at zero. Whether procurement integration and recaptured subcontracting margin close the gap is the entire justification, and it cannot be judged without the GS-bound revenue share. Synergy without the number is faith, not analysis. Exit is thin rather than absent — Singapore and Hong Kong track records put Japanese marine civil contractors, European geotechnical groups, and infrastructure funds in the plausible set — but the domestic strategic pool is narrow, so the defense belongs in the contract.
The desk’s design has four parts. First, deferral: a fixed price on conservative normalized EBITDA, with anything above it paid as contingent consideration tied to results and new orders over two to three years, so both sides can disbelieve each other’s future and still transact. Second, an escrow of 24 to 36 months released on completion and settlement of overseas projects rather than on the calendar, because funds have liquidation dates and an indemnity is worth only the counterparty that remains to pay it. Third, the limits of insurance: warranty and indemnity cover works for financial-statement and tax representations, but overseas construction claims and delay losses are usually excluded as known risks — and Dong-Ah has a tunnelling project under way in Singapore — so unbilled-receivable shortfalls and equipment residual values need a standalone special indemnity. Fourth, control: board composition, nomination of the finance chief, the annual capex ceiling, and approval of any single overseas project bid. GS E&C’s approx. KRW 119 billion loss at a water-treatment subsidiary’s overseas project is not an analogue for tunnelling risk, but it shows this organization has watched one overseas project move consolidated earnings, which is reason enough for an approval gate.
What Must Be True
1. The KRW 20 billion gap is depreciation, and depreciation is not a reinvestment obligation. Every number in this deal starts from an adjusted figure whose adjustments are undisclosed; debating the multiple before verifying the numerator is the wrong order. Falsifier: a reconciliation showing material normalization add-backs, or maintenance capex running near depreciation.
2. The synergy is net-new, and it stays. The integration story needs third-party revenue that survives the change of ownership. Falsifier: disclosure of a high GS-bound revenue share, or a visible decline in orders from competing general contractors after announcement.
3. The 2025 operating profit of KRW 18 billion is a run-rate, not a timing peak. The 88% rise from KRW 9.6 billion reflects a low base; beside KRW 22.9 billion in 2019, 2025 is a point on a recovery, not a breakout, and the cause of 2024’s weakness is unidentified. Construction accounting compounds this: percentage-of-completion assumptions move profit between periods, and whether a strong year is skill or timing cannot be known without comparing execution budgets to actual costs. Applying a multiple to a post-improvement high is the weakest joint in the sale logic. Falsifier: cost-overrun reversals or margin compression in the first post-closing prints.
Saemangeum touches that joint from outside. The deceleration case: the Saemangeum Development Agency’s 2027 budget proposal is KRW 222.2 billion, up 3.4%, and the KRW 120 billion for the new airport faces litigation at the Seoul Administrative Court. The acceleration case sits in the same document: corporate investment support rises 477% to KRW 26.5 billion, inter-regional roads receive KRW 143 billion, and Hyundai Motor Group reportedly signed an MOU in February 2026 for AI data center, hydrogen, and robotics facilities, with investment estimated by observers in the trillions of won. Both are true; the agency’s budget is one slice of a program spread across several ministries. The principle does not change: expectations and confirmed orders are different line items, and the base case takes confirmed backlog only. A second factor begins at closing — once the customer is the parent, profitability follows transfer-pricing policy rather than market rates, minority shareholders collide with it if the listing is retained, and related-party scrutiny becomes permanent. The risk an architect should fear most is not the one that explodes but the one that leaks forever.
The Architect’s Takeaway
What has been dissected is a KRW 300 billion transaction, but the point is not for the person holding KRW 300 billion. If you were building your own deal room with no capital and only the structural logic, four things transfer.
Split the claim. Crescendo put half into shares and half into bonds because the two claims behave differently when things go wrong. Without capital, refuse single-structure compensation: take a base fee and attach performance equity or a revenue share; separate fixed membership income from success fees. One activity, two claims, made by contract rather than money. One caution: this design did not get Crescendo out. A dual claim controls the size of a loss, not the speed of a recovery.
Make the earn-out your default. This deal has failed twice because the two sides believe different futures, and the architect’s tool for that is separation in time, not persuasion. When a client calls your rate too high, do not cut it; lower the base and add a performance band. A cut lowers your market price permanently; a conditional split preserves the price and moves only the risk.
Apply escrow logic to counterparties. Indemnities are worth only the party still standing to honor them. With organizations that dissolve by project, by funding round, or by an officeholder’s term, reduce the deferred share of your pay; with permanent institutions, accept deferral and grow the total. The counterparty’s life expectancy should set your payment schedule.
Distrust a number-one position without control. Dong-Ah is the top domestic operator in its methods, and someone else decides when those methods are deployed; irreplaceable skill allocated by others is a utilization-dependent asset, and decision rights are a separate acquisition from skill. The clue is in the cap table: Crescendo never held a majority, moving from 18–19% to 34%, and still runs the control process, because the largest single block among dispersed holders is the de facto decision. Find the point in your industry’s information flow where parties cannot see each other’s constraints, occupy it, make it recurring, and price it with two claims. And because a market price floors every negotiation once a plan is known, design quietly and disclose after the structure locks.
Syndicate’s Note
House Verdict: TRACK. The desk’s read is that this gap closes only through deferral: the seller’s floor is set by its own fund economics and 2019 entry structure, the buyer’s ceiling by after-tax return on capital with no verified synergy number, and negotiation moves neither. The falsifier is a signed agreement that pays the full ask as fixed, unconditional consideration — no contingent tranche, no project-linked escrow. The verdict classifies structural learning value; it is not an investment judgment.
Two things hold regardless of outcome. A domestic number-one position does not convert into a premium when the asset cannot decide when it works. And whoever does not ask what the KRW 20 billion between KRW 38 billion and KRW 18 billion actually is will pay what they believe to be six times, and never learn what they paid for.
FAQ
What is adjusted EBITDA, and why does the gap to operating profit matter in an acquisition?
Adjusted EBITDA is operating profit with depreciation, amortization, and items the seller classifies as non-recurring added back. The gap matters because the multiple is applied to it: if the add-backs are depreciation in a business that must replace its equipment, or normalization items that recur, the headline multiple understates the real price of the cash flow.
Why would a strategic buyer such as GS E&C use an earn-out for a long-time subcontractor?
Because the two sides can believe different futures. Contingent consideration fixes a price on conservative normalized earnings and pays the remainder only if the seller’s projected results and orders arrive, transferring proof risk to the seller. For a fund seller nearing liquidation, it also replaces an indemnity that may have no one left to honor it.
Does a number-one position in shield TBM justify a control premium for Dong-Ah Geological?
Not by itself. Dong-Ah’s methods are deployed when project owners and prime contractors award work, so its earnings follow utilization decided outside the company. A premium of 45–90% over the market-implied enterprise value, as the reported ask implies, has to be justified by verified synergy — which depends on the undisclosed share of revenue already earned from GS E&C.
This article is for informational purposes only and does not constitute investment advice, an offer, or a solicitation. Figures are based on public reports and may be estimates.
For more structural insights and deep-dive video breakdowns, visit Structure Syndicate on YouTube.