One market-research house values artificial intelligence in apparel at $35.71 billion by 2035, grown from a present it puts near six billion. It does not say how it counted either figure. The number is not a measurement; it is a mood, and the retail teams pasting it into strategy decks are quoting advertising they have mistaken for research.

Ask a second house and the floor drops. The Business Research Company sizes the same “AI in fashion” market at $1.75 billion for 2025, barely a third of the present the first firm reported. It then has that market growing at nearly forty percent a year, against the first firm’s twenty. They disagree not about the future but about the present, the one quantity a measurement exists to settle. Two houses, one named market, and no shared account of its size or its speed. Whatever is coming will not be clarified by people who cannot agree on the size of the room they stand in.

The engine under each of these numbers is identical. Take a modest present, apply a compounding rate almost no business sustains, and print whatever the far year returns; generative AI in retail, by one account, climbs from a billion dollars to twenty-one on a rate of just over thirty-five percent held over nine years. The rate is the product; the method that would let you audit it is the paragraph they decline to show. Few markets have grown at that pace for ten years, but the forecast assumes this one will, and dates the promise to a year no author will answer for.

Consider how a finding like this travels. The full report is sold, not published, and the summary goes out by newswire, the channel built for announcements rather than evidence. The headline reaches the reader; the arithmetic never leaves the invoice. Research asks to be checked; a press release asks only to be repeated.

They will quote the size of 2035 to the nearest ten thousand dollars and cannot place last year within three billion.

None of this is clumsiness. Bradford Cornell and Aswath Damodaran named the pattern in the Financial Analysts Journal, “the big market delusion,” after finding that the promise of a vast addressable market reliably inflates the value investors assign to the companies chasing it. They studied share prices, not sizing reports; the mechanism is the same. The larger the market you can assert, the more the assertion is worth to whoever paid for it. A forecast is a capital-raising instrument that happens to be shaped like a fact.

These reports pass for neutral third-party intelligence, the kind a board asks for before it signs a budget. Notice, then, which numbers stay sober. Grand View Research sizes the wider fashion-technology market at near six percent annual growth, a tenth of the rate any segment commands once the letters “AI” appear in the category name. The prefix does the lifting, because the prefix is what the vendor is selling. That is the figure in the trench coat: independent in posture, precise to the cent, produced by a trade whose customers are the firms it flatters. Read it as a brochure and it turns honest. Read it as strategy and you have let the salesman draft your plan.