Hyundai Rotem Company's (KRX:064350) defence division earned a 27.2% operating margin in the first quarter of 2026. The group as a whole earned 15.4%.
Defence was about 55% of revenue in that quarter. Run the arithmetic and the remaining 45% — rail rolling stock and the plant equipment business — contributed almost nothing to operating profit.
That is the company most investors are not looking at, and it is roughly half of it.
Hyundai Rotem builds railway rolling stock: metro cars, high-speed trainsets, locomotives, for Korean operators and for export. It also has a plant division supplying automated production equipment, largely to Hyundai Motor factories, which is the kind of related-party business a Korean group subsidiary tends to have.
Both are real businesses with real revenue. Neither has been reliably profitable.
The evidence is in the years before the tanks. Operating income was negative ₩196.2bn in FY2018 and negative ₩279.9bn in FY2019. Net losses were ₩308.0bn and ₩355.7bn. Retained earnings turned negative, reaching negative ₩156.6bn at the end of FY2019 and staying below zero until FY2022.
Revenue over that period was going backwards too: ₩3,309.1bn in FY2015 falling to ₩2,872.5bn in FY2021, six years of decline.
Those losses came from the rail and plant businesses, because there was nothing else. Export rolling stock contracts won on aggressive pricing ran into cost overruns, and under percentage-of-completion accounting the losses were recognised in full as soon as they became probable. It is the same mechanism that damaged Korea's shipyards in the same decade, applied to trains.
FY2025 revenue was ₩5,839.0bn, up 33.4%, with operating income of ₩1,005.6bn and a 17.2% margin. Net income was ₩770.5bn against a loss of ₩355.7bn six years earlier.
Essentially all of that improvement is the K2 tank export programme to Poland — the first contract, now delivered as to its initial 180 vehicles, and a second tranche reported at $6.5bn signed this year under a framework covering up to 1,000 tanks.
Which means the investment case rests on a single customer relationship, and the rest of the company is along for the ride.
There is a third category of revenue that gets almost no attention and probably deserves more.
On 25 June 2026 Hyundai Rotem signed a contract with Korea's Defense Acquisition Program Administration for outsourced maintenance of K-1 armoured recovery vehicles and K-1 bridge-laying tanks. Value: ₩198,523,000,000. Term: 25 June 2026 to 14 December 2029. Advance payment included, with progress billing. Zero-rated for value-added tax.
₩198.5bn is 3.4% of FY2025 revenue, spread across three and a half years — so roughly ₩57bn a year. Small. But sustainment work has qualities new production does not. The vehicles already exist, the scope is known, the customer is captive, and the revenue recurs for as long as the fleet is in service. Cost overrun risk on maintaining a vehicle you designed is a fraction of the risk on building a new one.
Every mature defence contractor eventually discovers that sustainment is worth more than production. Hyundai Rotem has an installed base of K-1 and K-2 vehicles in Korean service and, increasingly, in Polish service. The maintenance revenue from the Polish fleet has barely begun.
It also comes at the domestic margin. DAPA prices on a regulated basis, so this ₩198.5bn will not earn anything like 27%. That is the trade: lower margin, far lower risk, and a production line that stays warm between export tranches.
The practical consequence for an investor is about what happens after Poland.
Backlog was around $20.1bn at the first quarter of 2026, flat against the end of 2025 and well above the $12.7bn of end-2024. Flat backlog with rising revenue means the company is now delivering as fast as it wins. Analysts cut price targets in July on exactly that concern.
If a third Polish tranche arrives, the question is deferred. If it does not, then in two or three years Hyundai Rotem's revenue mix rotates back toward rail, plant and domestic defence — and the group margin rotates with it. The 27.2% defence margin will not survive a mix in which domestic and sustainment work dominate.
That is not a disaster scenario. It is the base case that ought to be modelled alongside the bull case, and it depends entirely on whether the rail and plant divisions can earn a mid-single-digit margin instead of zero.
The argument for the other half is better than its profit record suggests.
Rolling stock demand is genuinely growing. Rail electrification, urban metro expansion across Asia and the Middle East, and replacement cycles in developed markets all point the same way, and the competitive set — Alstom, Siemens Mobility, CRRC, Hitachi — is small. A Korean manufacturer with a functioning cost base is not obviously disadvantaged.
The losses of 2018 and 2019 were also specific: contracts priced to win share, not a structural inability to build trains. Companies do learn from that, and Hyundai Rotem's bid discipline has had six years to improve. The division has not lost money since.
There is a group argument too. The plant business exists partly to serve Hyundai Motor, which is a captive, creditworthy customer with a large ongoing factory automation programme. That is low-margin work with almost no receivable risk.
And keeping three divisions loaded is what allows the company to absorb the fixed costs of engineering capability between defence cycles. A pure tank manufacturer would be a far more volatile business.
Divisional operating margin for rail and plant, which the company discloses in its quarterly materials. The number to look for is not growth but the sign. Any consistent positive margin — even 3% — changes the valuation of half the company. Continued breakeven means the group is worth the defence division and nothing more.
Second, the size of the sustainment book. The company does not headline it, and the June K-1 contract only became visible because it crossed the disclosure threshold. If sustainment revenue is growing toward a tenth of the group, the earnings floor is much higher than the order book implies.
Third, a third Polish exercise. Everything else is second order until that is answered.
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.