Shein loses about $2.7 billion of paper value on its first morning in Hong Kong

Run the arithmetic backwards from the HK$44 print ^[1] and the level that 10% was measured against sits near HK$48.9 ^[7], which is to say Shein's syndicate priced the book within about a tenth of where public money was actually willing to own it. Then do the same exercise on Shanghai. CXMT's 466% first day means its shares went out at roughly 17.7 cents on the day-one dollar, about 5.7 times over ^[3][10]; Unitree's 629% means about 13.7 cents, or 7.3 times ^[4][11]. Graded on the one thing a bookbuild is supposed to get right, which is finding the clearing price, Hong Kong ran the accurate deal and the two mainland records were the errors.

The read being applied to this, that global money will capitalise Chinese memory chips and humanoid robots and will not capitalise Chinese clothes, is probably right in direction and close to untestable on the evidence to hand, because the Fortune Tech item that carries these numbers gives the pops and the market value without offer ranges, deal sizes, or free floats ^[6]. Shanghai and Hong Kong are not the same auction with different tickers.

The figure I keep going back to is the comp, or rather the more interesting version of the comp: a couple of ticks above Ralph Lauren ^[2] is where the market has parked a company whose weak open Fortune attributes to long-running investor doubts about the sustainability of the business ^[5]. That is a price, set over a morning by whoever happened to be liquid, and the roughly $2.7 billion that came off ^[9] is borne by the allocation rather than by the operating company, which keeps whatever the offering raised, a number the newsletter does not supply ^[6].

So, three ways this goes from here. Shein grinds back through HK$48.9, which takes 11.1% from HK$44 ^[8], and the break reads as a soft open rather than a verdict. Or CXMT and Unitree hand back most of 466% and 629% ^[3][4], and the summer gets relabelled an onshore liquidity event rather than a preference for hardware. Or Shein keeps sliding, and $24 billion ^[2] becomes the mark every private fast-fashion round is argued against. My view, held loosely: the discrimination is real but it is discrimination between pricing regimes at least as much as between business models, and the way to prove me wrong is to show me the three offer prices and the identity of the buyers, neither of which this source has.

The spread between Unitree's first day and Shein's runs to about 639 percentage points ^[12], and most of that gap was set by the offer prices rather than by the two companies.