Tendam lifted net profit by 26% last year and answered the result by committing to open more than 100 stores. The Spanish group behind Cortefiel, Springfield and Women’secret grew because more shoppers walked through a door — not because they found a better app. That is an awkward result for an industry that has spent three years insisting apparel discovery and checkout are moving to agents, feeds and conversational interfaces. The clearest data this season runs the other way.

The commerce-AI consensus holds that the store is a cost to be optimised away. In the version sold by vendors and echoed across much of the trade press, shoppers hand discovery to recommendation engines, styling agents and conversational checkout, and the shop sinks to a showroom, or to dead weight. Tendam’s year is a clean test of that pitch, and the pitch lost. The mechanism the forecast keeps missing is particular to clothing: most purchases turn on a question a screen cannot settle, which is whether the thing fits.

Tendam’s online channel did grow, which is the detail worth reading slowly. Digital sales rose 12.9% over the year. That leaves the channel at 17% of revenue, a share the consensus would log as proof of the migration it forecasts. The proof dissolves once you read what the share contains, because the online channel here does not compete with the stores; it runs on top of them.

Read the omnichannel figures and the word “online” begins to dissolve. More than half of online orders, 54.8%, are collected in a shop rather than shipped to a home. The great majority of online returns, 82.3%, come back through a counter rather than a courier. Store staff themselves ring up 28% of what the accounts record as digital sales. The pure form of e-commerce, in which a shopper orders on a screen, takes delivery at the door, returns by post and never touches the brand’s physical estate, is at Tendam a minority of a minority.

The store also quietly subsidises the website’s economics. Tendam’s online return rate sits near 26%, well below the 35% sector average Miquel cites, a gap that matters in a business where reverse logistics can erase a category’s margin. Returns fall when a customer can collect, inspect and, if it disappoints, hand an item straight back on the same trip, which a dense shop network makes easy. The website costs less to run because the shops are there to catch what it drops.

The profit did not come from a richer margin. Gross margin held roughly flat at 62.4%, so the jump was a volume story rather than a pricing one. It came from operating leverage: a denser network of shops pushing more sales through a broadly fixed cost base. Lower financing costs helped too, as Modaes reported, stretching an EBITDA gain of 9.4% into a net profit gain of 26%. Scale in physical retail still compounds, and Tendam is compounding it.

Follow the capital, and it points the same way. The investor that took control of Tendam in 2025 is not a software fund but Abu Dhabi’s 2PointZero, formerly Multiply Group, which acquired a controlling 67.91% stake in Tendam while CVC and PAI remained as minority shareholders, and is now bankrolling the expansion. The pace has sharply accelerated: Tendam opened about as many shops in the first quarter of this year as in the whole of the previous one. Capital that could have chased any digital thesis on the planet chose a brick-and-mortar apparel platform and a target of roughly 2,500 stores by 2030. Where the smart money goes is itself a finding, and it is going into leases.

The cheapest way to answer whether a garment fits is still to put it on — and no model has yet learned to be a fitting room.

The strongest objection is that Tendam proves nothing the market had not already granted Spain. Its base is Iberia and Mexico, store-dense and Spanish-speaking, plausibly insulated from the channel shift reshaping the United States and China. Its comparison year was soft. And the company is hardly anti-AI: chairman and chief executive Jaume Miquel describes “a firm and accelerated commitment to artificial intelligence where it generates the most value for us,” which a sceptic can read as Tendam confirming the consensus rather than refuting it. The thesis fails if physical-led growth is a regional quirk that cannot survive markets where discovery has already left the shop.

The objection does not survive the like-for-like numbers. Sales at established shops rose 6.1% over the year and 9.2% in the latest quarter, a measure that strips out new stores and currency, so the gain reflects deepening demand in shops Tendam already runs rather than the arithmetic of opening more. Tendam is also exporting the model rather than guarding a moat, into Romania, Hungary and the Balkans, with the United Kingdom under open discussion. A company does not commit a decade of capital to replicating a fortress it believes is purely local. As for the AI pledge, Miquel’s own criterion — “where it generates the most value for us” — in a retailer shaped like this points to the store’s machinery: clienteling, inventory and returns, not a channel meant to replace it. That distinction is the whole argument.

If fit remains the friction that closes a clothing sale, then the commerce-AI thesis has mislabelled a tool as a destination. The question was never whether machine intelligence enters apparel retail; it plainly does, in Tendam’s stockrooms and clienteling screens as much as anywhere. The question is whether it arrives as the shop or as the shop’s instrument, and Tendam has answered for itself by pointing its profit, and Abu Dhabi’s capital, at 100 more doors. The bill for getting that backwards falls on whoever built for a customer who was supposed to stop walking in.