Ecopro BM's revenue peaked at ₩6,900.9 billion in 2023. In 2025 it was ₩2,531.6 billion. Over the same period, the plant and equipment on its balance sheet went from ₩1,824.2 billion to ₩3,377.2 billion.
Ecopro BM Co., Ltd. (KOSDAQ:247540) makes high-nickel cathode powder, the most expensive part of a lithium-ion cell, for Samsung SDI and SK On. Its factories were commissioned during the 2022 and 2023 electric-vehicle boom. They arrived into 2025 and 2026.
The headline comparison above is unfair, and I want to deal with that before using it. Then I want to show that the fair version is not much better.
Cathode is sold on a pass-through formula. The maker buys lithium and nickel, converts them, and prices the output off the metal cost plus a conversion margin. When lithium collapsed after 2023, cathode revenue collapsed with it, whether or not a single tonne of volume was lost.
The gross margin proves it. In 2023, on ₩6,900.9 billion of revenue, gross profit was ₩293.4 billion. That is a 4.3% gross margin. In 2025, on ₩2,531.6 billion of revenue, gross profit was ₩269.8 billion, or 10.7%.
Read that again. Ecopro BM earned almost the same gross profit on a third of the revenue. Anyone treating the top-line drop as a demand collapse has the story wrong.
So use a denominator that is not distorted by metal prices. Compare operating income to the plant that produces it.
In 2023, operating income was ₩156.0 billion against ₩1,824.2 billion of property, plant and equipment. That is 8.6%. In 2025, operating income was ₩143.3 billion against ₩3,153.7 billion of plant, or 4.5%. Over the four quarters through June 2026, operating income was ₩131.0 billion against ₩3,377.2 billion. That is 3.9%.
The asset base nearly doubled and the profit it throws off did not move. That is the whole problem in one line, and it survives the pass-through correction.
Capex explains how the company got here. It spent ₩756.2 billion in 2023 and ₩1,023.8 billion in 2024, then ₩418.1 billion in 2025 and ₩161.7 billion in the first half of 2026. Roughly ₩2.36 trillion of plant went in over three and a half years, most of it committed when the order book looked different. The recent deceleration is the tell. Management has been slowing down, and reported as far back as mid-2024 that it might cut planned capacity outright.
Here is the measure I find hardest to argue away.
At the end of 2022, retained earnings were ₩432.2 billion. At June 30, 2026, after three and a half more years of operating, they were ₩277.8 billion.
Total equity over the same stretch rose from ₩1,488.2 billion to ₩2,096.5 billion. Equity grew ₩608 billion while retained earnings shrank ₩154 billion. The gap came from shareholders putting money in and from currency translation on the overseas plants, not from the business.
Look at the quarterly net income line to see why. The four quarters through June 2026 produced ₩79.0 billion of net income in total, and one of them, the June quarter, was a small loss. Finance costs are the reason: they ran roughly ₩110 billion over that year, against ₩131 billion of operating income. The plants earn money. The debt that built them takes most of it.
In 2025 Ecopro BM paid ₩60.0 billion of dividends. Net income that year was ₩91.6 billion, so the payout was about 65%.
Free cash flow in 2025 was negative. Operating cash flow was ₩214.7 billion and capex was ₩418.1 billion, a gap of ₩203.4 billion. Financing brought in ₩308.0 billion. The dividend was funded, in effect, by borrowing.
Twelve months later the company is asking shareholders for ₩1.2 trillion in a rights offering, with the first issue price set at ₩89,500. That price is 26% below the estimate published on June 30. A company that hands out ₩60 billion it did not generate, then asks for twenty times that amount at a discount, is telling you something about how it sequences decisions.
Cash conversion has not helped in 2026 either. Operating cash flow was negative in both the March and June quarters, at ₩88.9 billion and ₩72.2 billion. Inventories climbed to ₩791.9 billion from ₩602.8 billion at year-end, and receivables to ₩299.1 billion from ₩151.5 billion. Current liabilities of ₩2,121.7 billion now exceed current assets of ₩1,681.8 billion.
The bull case is straightforward and not stupid. Cathode plants take three or four years from decision to qualified output. A company that waits for demand to appear will miss the cycle entirely. Ecopro BM has a supply agreement with Samsung SDI reported at $34 billion, and the Hungarian plant that began production in December 2025 exists to serve European cells with European content. Volumes there are planned at around 10,000 tonnes this year and 30,000 tonnes in 2027, with a second line starting in September.
Underused plant is also cheap to fix. Utilisation is the single fastest lever on a fixed-cost business. If cell demand returns, the 3.9% return on plant goes up without another won of capex, which is the opposite of a company that has to spend to grow.
There is a valuation objection to all of it. The company's market capitalisation was ₩10.4 trillion at Friday's close. That is close to five times book value, and about three times the gross plant the argument above rests on. Idle capacity being an option does not tell you what the option is worth, and 4.9 times book is not a price that assumes a long wait.
Operating income per won of plant is the number I will keep tracking, quarter by quarter. It was 3.9% over the last year. If the Hungarian ramp and the second line lift it back toward the 8.6% of 2023 without further capex, the capacity thesis was right and the market is early.
The nearer marker is the rights offering. The final price gets struck later in the autumn, and the size of the discount will show what institutional buyers think this asset base is worth. After that, watch whether 2027 capex guidance stays near the ₩300 billion a year run rate the last twelve months imply. A company that keeps building into 3.9% returns is making the same decision twice.
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.