Kakao's Spinoff Bet: Why Wall Street's Favorite Korean Tech Play Is Losing Its Sparkle

The Stock Cut That Wasn’t About the Stock

Kiwoom Securities knocked 36.4 percent off Kakao’s target price this week. The move draws headlines, but the number that actually matters sits buried in a footnote: Kakao assigned its new AI subsidiary only 36.5 percent of combined asset value, while Kiwoom’s own calculation puts fair value at 56.4 percent. That gap—roughly 20 percentage points—is where the real story lives.

Kakao’s board announced a two-company split last month. Kakao X will carry the legacy business. Kakao AI gets the artificial intelligence portfolio. Investors called it the largest corporate restructuring in Korean tech this decade, and measured it against the Samsung Electronics spinoffs that reshaped the chaebol landscape in previous years. The comparison is apt only if you look at scale, not outcome.

The Math That Doesn’t Add Up

Kiwoom analyst Kim Jin-gu did something unusual: he priced the two entities separately and found they didn’t justify the transaction as structured.

Kakao X gets an implied value of 13.5 trillion won. Kakao AI gets 17.5 trillion won. Combined, that is 31 trillion won. Kiwoom’s own sum is essentially the same, but distributed differently—more into AI, less into the parent. The split, then, is not value creation. It is value redistribution, and potentially value destruction if the market penalizes Kakao X for being saddled with legacy costs.

Kim also cut this year’s revenue forecast by 8.3 percent and operating profit by 4.5 percent. Subscription projections for the KakaoTalk agent platform were slashed from a 10.9 percent paid user rate to 7.3 percent. Monthly subscription fees dropped from 135,000 won to 66,000 won. Agent-based advertising revenue fell from 4 billion won daily to 2.7 billion.

These are not trivial adjustments. They represent a fundamental reassessment of how much of Kakao’s future is actually tied to AI—and how much of that AI is genuinely transformative versus incremental.

The OpenAI Question

Kakao has courted OpenAI as a strategic partner. Kim’s report suggests the relationship is closer to a channel partnership than a technological alliance. ChatGPT runs inside KakaoTalk. That is distribution, not integration. The model itself lives elsewhere. The data flows both ways, but the strategic leverage tilts heavily toward whoever controls the underlying architecture.

This is a recurring pattern across the platform economy. Local super-apps have tried since 2022 to buy or bargain their way into AI leadership through partnerships with foreign model providers. The grants and licensing deals look impressive in press releases. The economics do not.

Kiwoom noted that Kakao’s vision of a “franchise backbone model”—an autonomous agent ecosystem that wraps multiple third-party services into a single subscription experience—remains unfinished. If that model never matures, the split becomes a bookkeeping exercise with real tax and regulatory consequences.

Who Wins, Who Loses

The obvious winner from the spinoff narrative is Kakao AI itself. If Kim is right that the AI unit deserves over half the combined value, the separate listing should, in theory, unlock that premium. New investors can buy pure-play Korean AI exposure without the baggage of KakaoTalk, gaming, and fintech. A dedicated valuation could attract funds that currently avoid the conglomerate structure.

Kakao X is the likely loser. It absorbs the legacy liabilities—the customer service obligations, the aging platform maintenance, the regulatory oversight that comes with being the dominant messaging app in Korea. Its growth story shifts from “platform plus AI” to “platform managing decline while AI company grows elsewhere.”

Employees are already restless. Labor unrest has accompanied the announcement. Workers at Kakao X face uncertainty about role transfers, compensation changes, and whether the split will create a two-tier employment system. That is a risk that does not show up in any discounted cash flow model.

Short sellers are active. The stock has fallen from the 110,000 won range Kiwoom once defended toward the 35,800 won level where it last traded—a move that erased roughly three-quarters of its target premium in a matter of weeks.

The Bigger Picture

Kakao is not an outlier. The spinoff impulse is spreading through Southeast Asian platform companies. Sea Limited has experimented with separating its gaming and commerce arms. GoTo in Indonesia is pursuing a similar restructuring. Grab has consistently argued that its verticals deserve independent valuations. The thesis is always the same: the market undervalues AI exposure when it is bundled inside a traditional platform company.

What this week’s Kakao developments suggest is that the thesis may be correct in theory and wrong in practice. Separating a business on paper does not create the capabilities that justify the separation. It requires genuine technological differentiation, defensible data moats, and a path to monetization that exists independently of the parent company’s distribution network.

OpenAI could provide some of that. But as long as the partnership looks like chatbot distribution rather than co-developed intelligence, the premium remains theoretical.

Kakao’s board appears to believe the market will reward the narrative. Kiwoom’s numbers say it will not. The stock price will decide which of them is right.

For now, the sharpest signal is not the target cut. It is the 20-point gap between what Kakao gave its AI subsidiary and what an independent analyst thinks it is worth. That gap measures the difference between a strategic bet and a hope.