Kakao splits into AI and legacy units in Korea's latest platform power play
Kakao's decision to split into KakaoAI and KakaoX signals how deeply the race to embed AI into every product layer is reshaping corporate governance — and why the market punished the stock anyway.
Kakao is betting its future on two separate companies
Kakao’s board approved a structural split on August 21, 2026 that separates its AI ambitions from its legacy businesses into two distinct entities: KakaoAI and KakaoX. The news sent shares tumbling more than 12 percent in intraday trading. Investors apparently weren’t convinced this was the right call — or at least not convinced it was the right call at this price.
Here is what actually happened and why it matters beyond South Korea.
The split itself
Kakao is using a
inhyeon bunbal
— a share-split restructuring — where existing shareholders receive proportional stakes in both new and surviving entities. KakaoAI receives 36 percent of the net asset book value; KakaoX takes the remaining 64 percent. It is not a spin-off in the Western sense where a parent retains controlling interest. Both companies will trade independently on the Korean exchange starting January 27, 2027, after a shareholder vote scheduled for December 17, 2026.
The leadership split is telling. Chung Sim-ah, currently Kakao’s CEO, takes the helm of KakaoAI. Kim Do-young, who runs Kakao Investment and oversees group investment strategy, leads KakaoX.
What each company actually does
KakaoAI is effectively KakaoTalk going standalone, but with an aggressive twist: the messaging app becomes an AI agent hub. The plan is to have the Kanana AI assistant learn user habits and then proactively connect services — search, recommendations, purchases, payments — all within the KakaoTalk conversation flow. The target is aggressive: a 50 percent increase in platform dwell time and 20 percent annual revenue growth, reaching 6 trillion won (
approximately $4.2 billion USD
) by 2030.
KakaoX inherits the cash cows and the investment arm. It owns KakaoBank, KakaoPay, KakaoPay Securities, Kakao Mobility, Kakao Entertainment, SM Entertainment, and Piccoma. It also controls roughly 4.1 trillion won in subsidiary investment capital plus an additional 2.3 trillion won from asset liquidation — totaling about 6.4 trillion won in deployable investment funds. Its target is more modest: 13.3 percent annual revenue growth through 2030.
Why this matters for the region
Kakao is not the first Korean platform giant to restructure around AI. SK Group has been building dedicated AI entities since 2023. Naver split its AI and commerce businesses earlier this decade. But Kakao’s move is the most complete structural break yet — and the one with the starkest implications for how Korean tech competes globally.
The logic is straightforward: AI demands speed, focus, and capital allocation that a conglomerate structure slows down. By separating the AI business from the financial and content businesses, Kakao hopes each entity can move faster, hire differently, and be valued on its own terms rather than averaged together in a single stock price.
The market’s reaction suggests doubt. A 12 percent drop on announcement day signals that investors see risk, not reward, in the short term. The questions are legitimate.
Who wins and who loses
KakaoAI wins if the AI agent strategy works. If users begin treating KakaoTalk as a daily command center — telling the AI to book rides, buy tickets, recommend restaurants, process payments — then the platform’s value reconfigures from a messaging app into something closer to a personal operating system. That is a far more valuable position than any app can claim on its own. The dwell-time target of a 50 percent increase is the critical metric. Without it, the agent economy never materializes.
KakaoX wins if it can deploy its 6.4 trillion won investment pool effectively. The fintech and mobility businesses are mature. The real question is whether the investment arm can find the next winners — whether in virtual assets, healthcare AI, or the next SM Entertainment-level cultural export. If Kim Do-young’s team picks well, KakaoX becomes a holding company with serious earning power. If it picks poorly, it becomes a slow-growth vehicle with too much capital and too few ideas.
KakaoTalk users stand to gain the most immediate benefit if the AI integration works smoothly. A conversational interface that actually connects services — not just recommends them — could redefine how Koreans interact with digital platforms. But if the AI agent feels clumsy or invasive, the backlash could damage the brand Kakao spent two decades building.
Competitors lose. Naver and Line (SoftBank) now face a Kakao that is sharper, more focused, and structurally optimized for AI speed. Both rivals have made similar moves, but Kakao’s move is the most drastic because it cleaves the company in half rather than creating internal AI divisions.
The unresolved risk
The split announcement followed years of stumbles in Kakao’s AI ambitions. Kakao Enterprise, the group’s B2B AI and cloud business, posted persistent losses and faced restructuring talks as early as 2023. The 2024 decision to develop a standalone AI platform app was an admission that embedding AI inside KakaoTalk alone was not fast enough. This split is the next logical step — but it is also an admission that previous strategies failed to deliver.
There is also the question of whether the 36/64 split ratio is fair. KakaoAI gets the dream and the target of 6 trillion won in revenue by 2030. KakaoX gets the real businesses and the investment capital. But KakaoAI’s valuation depends entirely on execution. If the agent strategy falters, the 36 percent stake becomes a question mark. If KakaoX underperforms, the 64 percent stake drags down the combined enterprise value.
What happens next
The shareholder vote on December 17 is the first real test. If institutional investors push back — and the stock drop suggests some already have — the plan could stall or be modified. Regulatory approval from the Financial Services Commission and the Korea Fair Trade Commission will also shape the timeline.
If the split completes as planned, January 27, 2027 marks the beginning of a new chapter for Korean tech. Two listed companies, one legacy, one speculative. The market will judge them separately. Kakao’s entire future rests on whether the AI side can convince investors that a messaging app can become an AI platform — and whether the rest of the company can fund that bet without starving itself.
No one outside Kakao’s boardroom knows if this will work. But the attempt itself reveals something important: Korea’s biggest tech platform now believes its only path forward is to split itself in two and hope the pieces are stronger apart.