Temu's Prices Were Never a Price. They Were a Subsidy Wearing a Price Tag.

In September 2022, a shopping app most Americans had never heard of appeared on the U.S. App Store, and within about six weeks it had bumped Amazon from the top of the free-app charts.¹ The pitch was a single irresistible word: cheap. A phone case for a dollar, a dress for four, a junk drawer's worth of gadgets for the price of one lunch. The story Temu told about that price was clean and flattering - direct from the factory, no middlemen, the savings passed to you. It was true. It was also nowhere near the whole truth.

The official story was that Temu was a cost innovation - a leaner supply chain that simply undercut everyone honestly. The real story is that the price on the tag was not a price at all. It was a subsidy in three layers, and Temu was paying you to find out it existed.

That is the thesis, and it matters because it changes what kind of company Temu is. A platform that wins on durable cost advantage is hard to kill. A platform that wins on stacked subsidies is only as safe as the subsidies - and subsidies can be legislated away in a sentence. Three pillars held the price down, and not one of them belonged to Temu.

The three crutches under a one-dollar phone case

The first crutch was the customs code. The U.S. de minimis exemption let any parcel valued under $800 enter duty-free, with minimal inspection. Temu's entire model - ship one cheap item at a time, straight from a Chinese factory to a U.S. doorstep - was engineered to live inside that loophole. The scale of it is staggering: de minimis shipments into the U.S. grew from roughly 134 million in 2015 to over 1.36 billion a year by 2024.⁷ Every one of those parcels skipped the tariff a container of the same goods would have paid. That was not efficiency. It was a regulatory gift Temu structured its logistics to harvest.

The second crutch was PDD's balance sheet. Temu did not merely price thin - it priced below cost. Goldman Sachs research estimated Temu lost roughly $6 per order in the U.S. in 2023 — one of several estimates in circulation, with other analyses putting the per-order loss considerably higher; one version puts customer-acquisition cost around $5 on a $39 order, down from about $16 on a $29 order the year before.¹⁰⁵ Read that improvement carefully: the loss was shrinking, but it was still a loss, deliberately funded by a parent company flush with profits from its Chinese marketplace. The third crutch was the megaphone - a blitz of advertising heavy enough to make Temu a household name in a year, paid for the same way.

Notice that only the last row is truly Temu's own. The factory-direct efficiency was real - but it was the smallest of the four levers, and the only one nobody could take away. The other three were borrowed, and the loud part of Temu's growth ran on the borrowed three. For a while, the borrowing worked beautifully.

For a while, the math looked like a miracle

The subsidy bought exactly what subsidies buy: velocity. Temu's parent reported total revenue up 94% year-over-year in the third quarter of 2023 and up 123% in the fourth, the acceleration explicitly tied to Temu's international push.² By September 2023 the platform claimed over 80 million U.S. active users and crossed 250 million cumulative downloads that year.⁴ Global gross merchandise value hit $15.1 billion in 2023, most of it in the back half.⁴ PDD's full-year 2024 revenue reached roughly $54 billion, up about 59% year-over-year.⁹ On a growth chart, this looks like one of the great consumer launches of the decade.

But velocity is not a moat. A blind look at the revenue curve hides the question that decides everything: what happens to the price when the crutches are kicked away? Temu never got to choose the answer. Washington chose it.

The day the loophole closed

On April 2, 2025, the White House signed an order ending the de minimis exemption for goods from China and Hong Kong, effective May 2. The terms were brutal to a parcel model: a duty of 30% of value or $25 per item, rising to $50 per item after June 1.⁶ On a one-dollar phone case, a $25 minimum duty is not a tax - it is the end of the business. Temu reacted within days, raising U.S. prices, halting direct shipments from China, and reining in the advertising blitz starting in late April.⁶ Then the second shoe dropped: a July 30 order eliminated de minimis globally, for every country, effective August 29 - pulling forward a deadline that had been set years out.⁷

The result was immediate and measurable. Temu's U.S. daily active users fell roughly 52% in May 2025 compared with March — more than half its daily audience gone in two months, according to Sensor Tower data reported by CNBC.¹¹ The platform stopped shipping to American consumers directly from Chinese factories and began the slow pivot to U.S.-based fulfillment, which is a different cost structure entirely.⁸ When the price floor lifted, the customers who had only ever come for the floor walked out the door. That is the tell of a subsidy, not a moat: when you stop paying people to be there, you find out how many were only there for the payment.

Wasn't this just an aggressive launch that's now maturing?

The fair objection is that every disruptor loses money to grow - Amazon did, Uber did - and Temu's per-order loss was already shrinking, from roughly $16 of acquisition cost on a $29 order to about $5 on a $39 order in a single year.⁵ By that read, Temu was simply walking up the normal curve from land-grab to profitability, and the tariffs are an unlucky bump, not a structural flaw. There is something to this: the loss really was narrowing, and the factory-direct logistics really are a durable edge no regulation can repeal.

But the comparison flatters Temu. Amazon and Uber subsidized to build something the law could not delete - a logistics network, a driver liquidity loop. Temu subsidized into a position whose largest single advantage was a customs exemption it did not own and could not defend. The improving loss-per-order is encouraging only if the price it was converging toward stays legal, and after May 2025 it did not. A drop of more than 50% in U.S. daily active users in two months is not the signature of a company maturing on schedule.¹¹ It is the signature of demand that was renting the discount and never bought the brand. The honest version: Temu built a real business on a real efficiency, then wrapped it in two subsidies it mistook for strategy - and the wrapper was load-bearing.

Temu's prices were one of the great magic tricks of modern retail, and like every magic trick, the wonder was in not seeing the wires. There were three of them. Two were cut by law in 2025, and the third - the only one Temu ever truly held - now has to do the work of all three.⁶⁷ The platform can rebuild a price around local sellers and domestic fulfillment, and it may yet survive on the genuine efficiency underneath. But it will be a different price, for a smaller crowd, at a slower pace. The dollar phone case was never the innovation. It was the bait - and the company is now learning, in real time and at real scale, what the world charges when the bait runs out.