Alibaba Promised to Break Itself Into Six. A Year Later It Was Quietly Gluing the Pieces Back On.

On the day Alibaba announced it would break itself into six, the market did the thing a market does when it likes a story: it bought. The U.S.-listed shares jumped double digits, more than 10% before the session was out.² The pitch was clean and irresistible — one of the world's most sprawling conglomerates, carved into focused pieces, each free to chase its own capital, its own valuation, its own IPO. A holding company that had grown too big to be priced correctly was going to let the market price the parts. Investors loved it because it promised to set a value free. One year later, the value was still locked in the vault, and Alibaba was quietly carrying the pieces back inside.

The story everyone filed under 'Alibaba splits into six independent companies' was wrong in the first verb. Alibaba did not split. It reorganized into a structure it labeled '1+6+N' — one holding company on top of six business groups — and that '1' never went anywhere.¹ The breakup was always a promise to test the parts against real markets later. The drama of the next twelve months is the story of that test arriving, and the parts failing it one by one.

The breakup was a bet that markets would reward the pieces: a separation agreement with the marriage left legally intact, and the crown jewel never put on the table

Here is the mechanism the headlines skipped. A '1+6+N' structure is not a divorce; it is a separation agreement with the marriage still legally intact. Alibaba Group remained the single holding company; the six groups — Cloud Intelligence, Taobao Tmall Commerce, Local Services, Cainiao Smart Logistics, Global Digital Commerce, and Digital Media and Entertainment — got their own CEOs and their own boards, but they remained subsidiaries.¹ The thing that would have made the breakup real was the next step: letting outside investors buy the parts directly. And the company was precise about which parts. Its own March 28 statement gave five of the six the flexibility 'to raise outside capital and potentially to seek its own IPO.'² The core domestic e-commerce engine — Taobao and Tmall — was never on that list. The crown jewel stayed inside; the satellites were sent out to be valued.

That is the whole logic of the bet. A conglomerate trades at a discount because the market cannot see, or trust, the value buried inside it. Spin the units off, let public investors set their prices in the open, and the sum of the visible parts should exceed the murky whole. The architect of that bet was Daniel Zhang, then Alibaba's chairman and CEO. The promise wasn't decentralization for its own sake. It was a public, market-tested re-pricing of a company the market had stopped believing in.

Then the market test arrived, and the parts failed it: where sanctions and a frozen IPO window did what no org chart could survive

The plan's centerpiece was the Cloud Intelligence Group. The board didn't just nominate it for an IPO — it approved a full spin-off, distributing the cloud business to shareholders as a stock dividend so it could stand as an independent listed company.³ This was the boldest version of the bet: not a partial float but a clean cut. And it ran straight into something no reorganization can engineer around. Washington's chip export restrictions meant the cloud unit couldn't reliably get the silicon a competitive cloud business runs on. In November 2023 Alibaba cancelled the spin-off outright, telling investors the restrictions 'may materially and adversely affect Cloud Intelligence Group's ability to offer products and services.'⁶ An independent cloud company you can't supply with chips isn't a value to unlock; it's a liability to disclose. The cancellation erased more than $20 billion of market value in the bargain.⁶

The logistics arm went the same way, from the opposite direction. Cainiao Smart Logistics had actually filed to go public in Hong Kong — the furthest any unit got toward proving the thesis. Then, in March 2024, Alibaba withdrew the application and proposed to buy out every minority shareholder for up to $3.75 billion, at $0.62 a share.⁷ Read that twice: the company that set out to sell pieces of Cainiao to the public was now offering to buy back the pieces it had already sold. Chairman Joseph Tsai pinned it on sluggish Asian market liquidity.⁷ The grocery chain Freshippo never even made it to the starting line, postponing its Hong Kong IPO in September 2023 on weak appetite for consumer stocks.⁸ Three units, three different exits — cancelled, withdrawn-and-absorbed, indefinitely deferred.

The personnel timeline rhymes with the strategic one. Zhang, who designed the structure and moved to personally run the unit that was supposed to be its proof of concept, was the first to come off. In June 2023 Alibaba announced he'd hand Group chairmanship to Tsai and the CEO role to Eddie Wu, staying on exclusively to lead Cloud and shepherd its spin-off.⁴ By September 10 he was gone from the cloud job too, leaving to start a private technology fund — the same day Wu also took the cloud helm.⁵ Within roughly six months of unveiling the plan, the man whose career was staked on it had no role in any of it. Two months after that, the spin-off he'd been kept on to deliver was dead.

Was it ever really a breakup — or just a holding company in a costume?: because a change you can switch off with a buyback and a withdrawn filing was only ever an option to change

The fair objection is that nothing here was a strategic blunder; it was bad luck and bad weather. Chip sanctions are an act of geopolitics, not a flaw in the org chart. Hong Kong's IPO window slammed shut on everyone, not just Alibaba. On that reading the company simply did the prudent thing — refusing to dump a chip-starved cloud business and a logistics unit into hostile markets at fire-sale prices. There is real truth in that, and the November filing's own logic concedes it: a full spin-off 'may not achieve the intended effect of shareholder value enhancement.'⁶ If the parts won't fetch a good price outside, keeping them inside is the rational move.

But that concession is exactly the point. A breakup whose every step is conditional on cooperative markets was never a breakup — it was an option to break up, exercisable only if conditions stayed perfect. And the one part that mattered most, the Taobao Tmall commerce engine, was deliberately fenced off from the whole exercise from day one.² The structure was built so the core never had to face the market test at all. When the satellites failed their tests and came home, what remained was the thing that had always remained: a single holding company with the same crown jewel inside it. The costume came off; the body underneath hadn't changed.

Alibaba set out to prove its parts were worth more in the open than its whole was in the dark. The cleverness of the plan was that the market would do the proving for it. The cruelty of it was that the market is also where the proving can fail — and when sanctions and a frozen IPO window turned the test against the company, it discovered it had built an elaborate mechanism it could simply switch off. The Cloud float was cancelled, the Cainiao shares were being bought back, the grocer never listed, and the architect had moved on to other things. The breakup didn't unlock the conglomerate. It just spent a year proving how hard the conglomerate is to take apart — and how badly its builders, when the moment came, didn't want to.