Kakao Pay Corp. (KRX:377300) has spent most of its listed life losing money, and the losses were not small: ₩57.5bn of operating loss in FY2024, ₩56.6bn in FY2023, ₩45.5bn in FY2022. FY2025 was the first profitable year, at ₩50.4bn. Then the first half of 2026 produced ₩90.8bn — nearly twice the whole of last year, in six months, against ₩13.7bn in the same half of 2025.
That's the headline, and it's been reported. The number nobody seems to be dating is further down the balance sheet. Retained earnings were negative ₩184.0bn at the end of 2024. They were negative ₩70.2bn at the end of June. The deficit is closing at roughly ₩40bn a quarter now, which puts it at zero somewhere around the turn of the year.
In Korea a company pays dividends out of distributable profit, 배당가능이익, which is calculated from net assets after deducting paid-in capital and the statutory reserves. A carried-forward deficit, 결손금, sits in front of that calculation. Until it's cleared, ordinary dividends are effectively off the table. Kakao Pay has never paid one.
A US reader will find this stricter than what they're used to. State corporation law generally lets a company with an accumulated deficit still pay out of current-year earnings or surplus, subject to a solvency test, so an American fintech turning profitable can declare a dividend the same year if the board wants to. Korean companies can also clear a deficit by resolving to offset it against capital reserves rather than waiting to earn it back, and Kakao Pay could do that at a shareholder meeting. It hasn't. Which means the arithmetic on the earnings line is, for now, the thing that decides when this company becomes capable of returning cash at all.
That reframes what to argue about. Not whether Q3 sets another record, but whether the FY2026 annual report closes with retained earnings above zero and what the board says about it in March.
On June 23 the board resolved to acquire the remaining 2,914,652 shares of Kakao Pay Securities Corp. for ₩172,981,681,548 in cash, taking ownership from 72.93% to 100%, with completion set for July 20. The filing puts the price at 8.87% of Kakao Pay's consolidated equity and 3.24% of its total assets. The stated purpose is improving governance inside the group to raise management efficiency, which is the kind of sentence that means the deal is about control rather than about anything operational.
The financial summary attached to the filing is the useful part. Kakao Pay Securities earned ₩41.0bn of net income in FY2025 on ₩242.0bn of revenue, after losing ₩26.1bn in FY2024 and ₩51.7bn in FY2023. Its equity was ₩204.6bn. So the parent paid ₩173.0bn for 27.07% of a company with ₩204.6bn of book, implying about ₩638.9bn for the whole thing, or 3.1 times book and roughly 15.6 times last year's earnings.
For a brokerage that has been profitable for exactly one year, 3.1 times book is not cheap. Korean securities firms more often trade below book. What the price says is that Kakao Pay's management believes the brokerage's earnings are durable, and that they'd rather own the upside than share it. Set that ₩41.0bn of FY2025 net income against the parent's own ₩55.7bn and the subsidiary was most of the group's profit last year. Buying in the minority is the most expensive way to express confidence, and they did it in cash.
Total payment volume in Q2 was ₩54.2tn, up 20% year on year, with revenue-generating volume at ₩15.7tn, up 19%, per the company's Q2 materials. Revenue in the same quarter was ₩335.1bn, up 41%. For the half, revenue of ₩635.4bn against ₩450.2bn is the same 41%.
Revenue growing at twice the rate of volume means the company is earning more per transaction, and that only happens through mix. The company says digital finance — insurance, brokerage, lending — passed 52% of revenue in Q2. Payments is a thin-margin business anywhere, and Korea's is thinner than most because the merchant fee structure is regulated. The bull case for this stock was always that Kakao Pay could convert people who tap their phone at a convenience store into people who buy an insurance policy inside the same app. The 2026 numbers are the first period where the conversion shows up in the operating line rather than in a slide.
Anyone screening this company on leverage will get the wrong answer. Total liabilities were ₩4,822.7bn at the end of June against ₩2,038.9bn of equity, and current liabilities alone were ₩4,703.7bn. Almost none of that is borrowing. It's customer float and settlement payables — prepaid balances, money in transit, brokerage client deposits — matched by ₩6,348.0bn of current assets on the other side.
The same distortion runs through the cash flow statement. Operating cash flow in Q2 was negative ₩112.3bn even though the company earned ₩49.6bn, and Q3 2025 was negative ₩133.3bn on a profitable quarter too. Those swings are settlement timing, not burn. Judge this company on the income statement and on the deficit line; the cash flow statement of a payments processor tells you mostly about what day of the week the quarter ended.
The valuation still assumes a lot. At ₩45,900 the market cap is ₩6.21tn, which is about 37 times first-half earnings annualised and 3.05 times book. Trailing FY2025 earnings put it at 111 times. The profit inflection is steep enough that the multiple compresses fast if it holds, and doesn't if it doesn't. Two quarters is not a trend, and the brokerage business that drove it is levered to retail trading activity, which is cyclical everywhere and unusually so in Korea.
Ant Group's stake, reported at about 27%, is the structural overhang. It has been there since before the listing, and any sale would be a large block into a stock that trades 35% below its 52-week high. Nothing suggests a sale is imminent. But a foreign strategic holder of that size in a regulated Korean financial company is a permanent input into how this trades, and it interacts with an unsettled policy environment — the company itself flags regulatory uncertainty around stablecoins and AI services.
And the honest bear point on the securities buyout: paying 3.1 times book for a subsidiary you already controlled adds no revenue, no customers, and no capability. It converts ₩173.0bn of cash into a larger share of profits the company was already consolidating. If the brokerage's 2025 result was the cycle peak, that price will look poor.
The FY2026 annual report in March, and one line in it: retained earnings. If it prints positive, this becomes a company that can legally pay a dividend for the first time, and whatever the board says next about capital return is the first such statement in its history. If the deficit is still there, the earnings ramp slowed and the March commentary will say why.
Before then, the Q3 report in late October gives the first quarter with Kakao Pay Securities consolidated at 100%, so the minority interest line disappears and reported net income attributable to the parent should step up mechanically. Worth separating that from underlying growth when the number lands.
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.