Alibaba's Real Money Machine Is an Ad Network Wearing a Mall Costume

Picture a billion-square-foot mall where the landlord owns no stores. He doesn't buy the dresses, doesn't stock the phones, doesn't ring up a single sale. He owns one thing: the directory at the entrance and the spotlights over each stall. Want to be on the first page? Want the bright light? Pay him. That is the heart of Alibaba's China commerce business - the part the company's own filings call customer management revenue, charged on clicks, impressions, time slots, and completed sales.² On RMB 941 billion of total revenue in fiscal 2024¹, the quiet truth is that the real machine isn't a retailer selling you things. It's an attention auction wearing a mall costume.

The official story is that Alibaba is 'the Amazon of China' - a giant that buys low and sells high. The more useful correction: its core engine doesn't touch the goods at all. It rents visibility to the merchants who do, and lets them carry the inventory, the markdowns, and the risk.

The profit hides in the search bar, not the shelf

Here is the mechanism, worked down. On Taobao and Tmall, millions of merchants compete for the same scarce thing: a shopper's eyeballs. Alibaba doesn't decide what gets sold - it decides what gets seen first. Merchants bid for that placement, and the fees they pay are billed on a cost-per-click, cost-per-impression, time, or cost-per-sale basis.² Notice what that is: it's an advertising business, structurally closer to Google search than to a department store. The shopper never sees a price tag from Alibaba, because Alibaba isn't the seller. It collects a fee for organizing demand, and demand is the one input that doesn't sit in a warehouse depreciating. When customer management revenue at Taobao and Tmall grew 9% in the December 2024 quarter, the company attributed it to two levers and two only: more GMV flowing through the marketplace, and a better take rate.⁷ No truckloads. No procurement. Just a thicker auction.

The 'zero inventory' line is a myth - and the filings say so

Now the part the tidy 'pure marketplace' story leaves out, because it's inconvenient. Open Alibaba's FY2024 results and there is a second, very un-asset-light line sitting right next to the auction: 'direct sales and others,' RMB 110.4 billion - about US$15.3 billion - up 6% year-over-year, driven by consumer electronics and appliances.⁴ That is not a fee on someone else's goods. That is Alibaba acting as principal, taking title, and booking revenue gross. Phones and refrigerators it actually buys and actually sells. The marketplace-not-inventory thesis is directionally right about where the profit lives, and flatly wrong as a description of the whole company.

And the wind-down is real, not retroactive denial. By the December 2024 quarter, direct sales and others had fallen to RMB 28.7 billion, down 9% year-over-year, which the company pinned on a 'planned reduction of certain direct sales businesses.'⁵ Read that carefully: you can only plan to reduce a business that exists at scale. Alibaba isn't telling you it never held inventory. It's telling you it's choosing to hold less of it - steering back toward the higher-margin auction it would rather run.

It owns the warehouses too - and just bought more of them

If inventory were the only crack in the asset-light story, you could call it a rounding error. Logistics closes the case. Cainiao isn't an arms-length partner Alibaba books a profit from - it's a fully consolidated subsidiary, roughly RMB 99 billion of revenue in fiscal 2024.⁶⁸ That is a logistics company the size of a midcap, sitting inside the books, consuming capital the way logistics always does. A pure marketplace would never own this. And the direction is the tell: in March 2024 Alibaba announced a plan to buy out Cainiao's remaining minority shareholders entirely.⁶ When you believe in asset-light, you sell the warehouses. Alibaba is buying the rest of them.

So is the marketplace thesis just wrong?

The fair objection is that all of this demolishes the headline: if Alibaba sells goods and owns logistics, calling it a marketplace is marketing. But that conflates two different questions - where the revenue is, and where the profit engine is. The direct-sales and logistics lines are large, capital-hungry, and lower-margin; they are exactly the businesses Alibaba is trimming. The customer management auction is the one it's protecting and growing, because a fee on someone else's GMV needs no warehouse and scales like software. The honest counter, though, is that this engine is not a fixed lever. Alibaba's own filings show the take rate fell in the March 2024 quarter, because lower-monetization Taobao GMV grew faster than higher-monetization Tmall, and new low-monetization models were added.³ The toll can be cut - by the company chasing volume, by competitors, by regulators. The marketplace is the best business Alibaba runs. It just isn't a fixed-margin annuity, and the company has spent more than a decade hedging that bet with the very inventory and warehouses the legend says it never touches.

Alibaba makes its money the way a landlord with floodlights does - serenely indifferent to whether the dress sells, as long as the merchant paid to stand under the light. That is the engine, and it's a beautiful one. But strip the costume off and the 'asset-light marketplace' is also buying refrigerators, owning warehouses, and writing checks to consolidate a logistics empire. The real Alibaba isn't a pure marketplace or a pure retailer. It's a company that learned the spotlight is the best business it has - and then, just in case, bought the building, the loading dock, and a fair amount of the merchandise too.