Hyundai Rotem Company (KRX:064350) held ₩908,414,081,000 of cash at the end of December 2025. Three months later it held ₩2,681,697,361,000.
Operating cash flow in the first quarter of 2026 was ₩2,156,076,284,000. For a company whose entire quarterly revenue was ₩1,457.5bn.
Cash flow of one and a half times revenue in a quarter does not come from selling things. It comes from being paid in advance for things not yet built.
The sequence is visible across the balance sheet. Current liabilities jumped from ₩3,222.5bn at the end of September 2025 to ₩5,776.4bn at the end of December — an increase of ₩2,553.8bn, while cash rose only ₩234.9bn over the same three months. The obligation was recognised before the money arrived.
Then in the first quarter of 2026 the money arrived.
The cause is the second Poland contract for K2 main battle tanks, reported at $6.5bn, the second tranche exercised under a framework covering up to 1,000 vehicles. Large defence export contracts carry substantial down payments — the customer funds the production ramp because the manufacturer cannot.
Total assets went from ₩6,489.7bn at the end of September 2025 to ₩9,318.0bn at the end of December. The company grew by 44% in a quarter without acquiring anything.
This is the part that matters, and the balance sheet says it plainly.
Current liabilities at 30 June were ₩5,181.7bn. Total equity was ₩3,369.8bn. The liabilities are half again as large as the equity, and the great majority of them are contract liabilities — payments received against tanks Hyundai Rotem has promised to deliver over the coming years.
Cash of ₩2,527.3bn equals about 75% of total equity and 28% of total assets. If you netted the contract liabilities against it, the company's own free cash would be a fraction of the headline figure.
None of this is unusual for defence manufacturing and none of it is a risk in the ordinary sense — the counterparty is a NATO government and the contract is signed. It does mean that anyone looking at ₩2.5tn of cash and thinking about buybacks or acquisitions is looking at a float, not a war chest.
The second quarter demonstrated it. Operating cash flow across the first half was ₩1,544.5bn on the cumulative basis Korean interim statements use, against ₩2,156.1bn at the first quarter. So the second quarter consumed roughly ₩612bn. The company has started building, and the money is going out at the rate the work requires.
Against that background, the dividend record.
Dividends paid were zero in FY2023, ₩10.9bn in FY2024 and ₩21.8bn in FY2025. Net income in FY2025 was ₩770.5bn. That is a payout ratio of 2.8%.
The first half of 2026 shows ₩65.5bn — three times the whole of last year, against first-half net income of ₩389.0bn, or about 17%.
So the direction is right and the level is still very low. A company earning ₩770bn a year, holding ₩2.5tn of cash, inside the Hyundai Motor Group, distributing under 3% of its profit, is either being extremely prudent about a contract cycle it does not yet trust, or has not thought hard about what to do with the money.
Capital spending suggests it is not investment. Purchases of property, plant and equipment were ₩129.2bn in FY2025 and ₩32.2bn across the first half of 2026, against a plant balance of ₩1,832.4bn. Investing activities were a net inflow of ₩119.3bn in the half. This company is not building anything with its cash.
One item that distorts the book value and deserves a flag.
In the second quarter of 2025, property, plant and equipment jumped from ₩1,311.0bn to ₩1,775.0bn — an increase of ₩464.0bn in three months, against capital spending far too small to explain it. In the same quarter, total comprehensive income was ₩514.9bn while net income was ₩189.5bn, a gap of ₩325.4bn. Total equity rose ₩515.0bn.
That pattern is a revaluation of land or buildings, taken through other comprehensive income rather than the income statement. It is entirely legitimate under IFRS and Korean industrial companies with old industrial sites do it periodically.
The consequence is that anyone computing book value per share, or a price-to-book multiple, using figures after the second quarter of 2025 is including roughly ₩325bn of asset revaluation that no operating performance produced. Retained earnings of ₩1,755.4bn are unaffected; total equity of ₩3,369.8bn is not.
There is a serious argument for the conservatism, and it is about what the cash is for.
Hyundai Rotem is at the start of delivering a $6.5bn contract, in a business where cost overruns are recognised immediately and in full. Holding customer advances rather than distributing them is exactly what a prudent contractor does, because if the programme costs more than planned, that money is what absorbs it.
The company's own history supports the caution. It reported operating losses in FY2018 and FY2019, of ₩196.2bn and ₩279.9bn, and carried negative retained earnings as recently as FY2021. Management that lived through that will be slow to hand cash back.
And the mix is turning against margins in the near term — operating margin has fallen for five consecutive quarters as the new Polish contract enters its early, lower-margin phase. Distributing aggressively into that would be poor timing.
The FY2026 dividend, declared with full-year results. Payout went from 2.8% toward 17% in the space of six months. A full-year figure above 20% would say the board has accepted that the cash cycle is manageable and shareholders should participate. A retreat back toward 10% would say it is still treating every won as the customer's.
Second, the contract liability balance disclosed in the notes, separately from other current liabilities. That single figure tells you how much of the ₩2.5tn is genuinely spoken for. Without it, the cash balance is uninterpretable.
Third, whether operating cash flow stays negative through the delivery phase. Two more quarters like the second — an outflow of ₩612bn — would consume the entire advance within about a year, and at that point the question of what to do with the cash answers itself.
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.