Kakao Corp. (KRX:035720) grew revenue 45.1% in FY2021. Then 15.0% in FY2022, 11.1% in FY2023, 4.1% in FY2024 and 3.0% in FY2025, to ₩8,099,147,815,086.
Five consecutive years of deceleration, ending in low single digits at a company that is still described, and valued, as a growth platform.
Over the same period the operating margin did the opposite. It was 9.9% in FY2021, fell to 8.4%, then 6.4% and 6.3% in FY2024, and has since recovered to 9.0% in FY2025 and 12.1% across the first half of 2026.
The second quarter of 2026 produced record revenue of ₩2,098.5bn and record operating income of ₩277.0bn. The record profit is real. It came from the cost line.
Operating income rose from ₩495.3bn in FY2024 to ₩732.0bn in FY2025 — a 47.8% increase — on revenue growth of 3.0%. Nothing about that ratio comes from selling more.
Kakao spent 2024 and 2025 reducing the number of affiliates, closing or selling businesses, and cutting cost. One Korean headline summarised it as losing weight by starving, which is unkind and roughly accurate. The company shed operations rather than growing into its overhead.
That is a legitimate response to the problem it had. It is also a strategy with an arithmetic limit: a company can improve its margin from 6.3% to 12.1%, and it can do that once.
The segment detail shows where the remaining growth sits.
Second-quarter platform revenue was ₩1,230.3bn, up 17% year on year. Within it, TalkBiz — the advertising and commerce built into KakaoTalk — was ₩643.2bn, up 12%, and the advertising and subscription portion of that was ₩399.9bn, up 14%.
Platform is therefore about 59% of group revenue and growing at 17%. Group revenue grew 9%. Subtract and the remaining 41% — content: games, music, web comics and stories — was flat at best and more likely shrinking.
That matters because the content businesses were the acquisition-driven part of the group, and they are the part being sold. The second quarter carried a ₩180.8bn loss from discontinued operations arising from deconsolidations and stake sales.
So the company is becoming a more concentrated advertising and commerce business attached to a messaging app. That is a better business than the sprawl. It is also a smaller one, and its growth rate is the group's growth rate now.
Kakao's answer to the growth question is artificial intelligence, and the company has been specific about the timeline in a way that deserves credit.
It targets more than 10m monthly active users of AI services inside KakaoTalk by the end of 2026. It expects meaningful AI revenue to begin in 2027 and to expand faster through 2028. And in the value-up plan approved on 21 August, the entity provisionally named Kakao AI is targeted at more than ₩6tn of revenue by 2030 with an operating margin above 30% and a return on equity above 25%.
Read that timeline plainly. Users this year, revenue next year, scale the year after, targets in 2030.
Companies that put dates on things should be held to them, and this is a company that has earned scepticism — the affiliate expansion of 2020-22 came with its own confident projections. But specificity is better than vagueness, and December's user number is the first checkable milestone.
The gap between now and then is what an investor is being asked to fund. Kakao held ₩8,178.1bn of cash at 30 June, which is more than enough to fund it.
One technical point for anyone reconciling figures, because two sets are circulating.
Kakao reported second-quarter revenue up 9% and operating profit up 36% year on year. The DART interim series shows second-quarter 2025 revenue of ₩2,028.3bn and operating income of ₩185.9bn, which against 2026's ₩2,098.5bn and ₩277.0bn would imply growth of 3.5% and 49%.
Both are right. The company's growth rates use prior-year comparatives restated to exclude the businesses since deconsolidated; the interim filing series carries the figures as originally reported. With this much disposal activity, that difference is material — it is the gap between 3.5% and 9% revenue growth.
Use the company's restated basis for growth rates, and the filed statements for levels. Do not mix them.
The sympathetic view is strong and probably correct.
Kakao in 2022 was a company with a 6% operating margin, more than a hundred affiliates, and a political problem. It had bought growth it could not integrate and had drawn genuine public hostility over how much of Korean daily life one platform touched. Fixing the cost base and the affiliate count was not optional; it was the condition for being left alone.
Doubling the margin while shedding businesses is a considerable operational achievement, and it has been done without damaging the core: platform revenue growing 17% says the advertising business inside KakaoTalk is healthier than it was, not weaker.
There is also an argument that the deceleration was inevitable regardless. Kakao's addressable market is a country of 52m people, essentially all of whom already use the app. Growth from here has to come from monetising the same users better, which is precisely what a 14% increase in advertising and subscription revenue represents.
Platform revenue growth in the second half. It ran 17% in the second quarter. Holding above 15% while content shrinks would mean group revenue growth accelerates on its own as the disposals annualise out. Dropping to single digits would mean the core has slowed too, and the margin story would be all there is.
Second, the 10m AI user figure in December. It is dated, specific, and the company chose it.
Third, whether operating margin holds above 12% once the cost programme is complete. The gains from FY2024's 6.3% came from removing things. There is a level below which a platform company cannot cut without damaging the product, and Kakao is closer to it than it was two years ago.
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.