Shein’s Hong Kong listing document offers public investors a base of 273 million annual active customers. The company cleared the exchange’s listing committee on 17 July and published its post-hearing information pack on 26 July. Nothing in that document reconciles the customer count with the fact that the parcel regime which produced it no longer operates in the United States or the European Union. Investors are being asked to underwrite a cohort whose unit economics were rewritten after it was bought.

Washington closed the duty-free parcel before Brussels did, moving months ahead of an inventory Shein could not relocate as quickly. Executive Order 14256 ended de minimis entry for goods of Chinese and Hong Kong origin on 2 May 2025, and Executive Order 14324 extended the suspension to every origin on 29 August. Two orders, four months apart, closed the loophole first for Shein’s largest single market and then for every market still using it.

Brussels took longer and chose a blunter instrument. Since 1 July the EU has applied a customs duty of €3 per item on parcels valued under €150, agreed by the Council in December. That duty lands on a market Shein cannot treat as secondary: Europe and the United States together carried a little over 57 percent of Shein’s 2025 revenue.

The base in the filing was assembled before either duty existed. Shein’s annual active count stood at 227 million when 2024 closed, which means five of every six buyers in the headline figure had transacted before the United States charged anything on a Chinese parcel. The European levy is four weeks old, so no part of the reported cohort has been observed under it at all. A customer count compounding above 21 percent a year through 2025 is a measurement of the old regime rather than evidence about the new one.

The objection circulating about this listing is a valuation objection, and it is the smaller one. Reuters found investors questioning whether Shein can justify the $40 billion to $50 billion it is seeking against a 2022 private mark near $98 billion, with Shen Meng of Chanson & Co telling the agency he expects no substantial uplift on the last funding round. That is an argument about the multiple. The cohort argument is about what sits underneath it: whether the customer base is a durable asset or a census of a subsidy.

Shein bought this cohort at one price and now has to hold it at another. Marketing ran at 10.7 percent of net revenue in 2024, a ratio that had barely moved since 2023. It climbed to 14.8 percent across all of 2025, then to 15.8 percent in the first quarter of 2026. Most of that increase happened over the course of last year, not suddenly in the three months just disclosed. Five extra points of revenue are going into holding a base that used to renew itself far more cheaply, which is what happens when the price of acquiring a customer and the price of keeping one stop being the same number.

The cost of moving the box has climbed in every year since the duties began. Fulfilment took 42.1 percent of net revenue in 2023 and reached 47.7 percent in the first quarter of 2026, which makes it the largest single claim on every dollar Shein collects. Shein’s own account, relayed in its disclosures, is unadorned: after the removal of duty-free treatment for low-value parcels, fulfilment cost per order has continued to increase.

Shein’s answer was to raise prices, and for two years it worked. Gross margin widened by almost eight points between 2023 and 2025, the arithmetic of a toll handed to the buyer rather than absorbed by the seller. Then the market with a full year of duty behind it stopped growing: American revenue fell 14.3 percent year on year in the first quarter, on a filing that also disclosed duties of 10 to 87.5 percent on Chinese-origin goods sold into the US. A margin problem being solved does not look like that; a cohort declining to pay does.

A cohort acquired under a subsidy is a fact about the subsidy until it has been tested without one.

The strongest case against this reading is that frequency has not moved. Orders per customer have held close to four a year in every period disclosed, including 2025, when the American duty landed mid-year and Shein still added buyers; the count reached 281 million by March 2026. For the argument here to fail, that stability has to hold specifically inside the two taxed regions rather than on the blended average, and Shein has to convert enough parcels into domestic shipments to take the duty out of the calculation. Both conditions are testable, and one of them is already returning results.

The blended average is where the divergence hides. The United States is the only market with four full quarters of duty behind it and the only large market shrinking; group revenue still grew 1.1 percent in the first quarter because the untaxed regions carried it. Domestication is the harder condition, because products held in central warehouses in China accounted for more than 90 percent of net revenue in 2025, and that is not a footprint which relocates inside a listing cycle. The instruments are also still multiplying: as we reported earlier today, Section 338 of the Tariff Act of 1930 has been revived to set duties by proclamation with no injury finding at all.

None of this makes the offer unbuyable; it makes the disclosure incomplete in one specific place. If orders per customer holds in Europe through the second half and the American decline flattens, the duty was a margin event and the customer base was real. If frequency slips in the taxed regions while the untaxed ones carry the average, then subscribers will have bought a cohort at a price set by the era that produced it and held it through the era that reprices it. The prospectus reports the first outcome as history and does not price the second, which leaves the choice with whoever signs the cheque.