The Watson Weekly is sponsored by Avalara — the agentic AI platform automating global tax and compliance for leading eCommerce brands with simple plans, transparent pricing, and complete coverage across Shopify, Stripe, WooCommerce, BigCommerce, and more.

H&M's resale operations, covering Sellpy and brands including COS, accounted for roughly 0.8% of company revenue in 2025, by H&M's own disclosure. Any claim about secondhand entering the mainstream has to survive that number first.

Fast fashion works because one style gets bought in enormous quantity and pushed through a system built for repetition. Secondhand inventory is one of one. Sourcing, inspection, grading, pricing, and merchandising all happen at the individual garment level, which means labor per unit stays flat no matter how much volume gets added. That may be why the pop-up remains the preferred format. A pop-up is never asked to prove it scales.

The forecasts arrive anyway. ThredUp commissioned a report putting the US secondhand apparel market at $78.8 billion by 2030, a number worth reading with the sponsor in mind, and one that bundles peer-to-peer selling, consignment, thrift, and platform GMV into a total no single retailer can actually address.

The brands that built their identity on sustainability are being called out in public right now. Everlane sold to Shein. Allbirds shut its apparel stores, reinvented itself as an AI company, and the stock rose on the news. Reformation went public last week carrying a preloved section with Prada and Bebe alongside its own product, and that one reads differently from the others, because Reformation started in upcycled vintage and the section is closer to where the company came from than to a pilot.

Vishal Agrawal at Georgetown argues the actual fix is designing clothes that outlast a trend cycle, and that reuse deserves more retailer attention than textile recycling. The chemistry supports him. Recycling struggles with blends, and most garments are blends: polyester with cotton, rayon, nylon, then zippers and buttons that have to come out before anything else happens. Nobody has solved that at cost in a decade of trying.

Regulation has moved faster. France passed legislation to fine apparel companies for churning out high volumes of cheap clothing, and California enacted a law requiring retailers to manage the disposal of clothing waste. Those are cost inputs rather than reputational ones, and cost inputs tend to change merchandising behavior on a shorter timeline than coverage does.

The question I'd put to any retailer running one of these programs is what revenue share it would take before secondhand stops being a marketing surface and starts being inventory strategy. At 0.8%, it's a rounding error set against the volume H&M puts into the world every quarter.

It's also a line item that didn't exist a few years ago and is growing off a base of nothing. Nobody has said which of those two readings they're actually managing toward, and I suspect that ambiguity is doing useful work for them.

The July 31 filing shows $15B in the first quarter, $13.7 billion in the second, and the remaining $21.3B after June 30. February's terms tied that back half to an IPO or to milestones neither party has described. OpenAI has not gone public. Something else cleared.

Amazon told investors the money is out the door. It did not tell them what OpenAI had to do to get it. That gap will matter more in the next filing than in this one.