The profit inside sustainable fashion

On September 3rd, at Hong Kong's Palace Museum, the organisers of Fashion Summit (HK) 2026 — billed as "Asia's leading sustainable fashion event", under the theme "Code the Fashion Forward" — promised an audience that "data-led strategies and AI" would build "agile, demand-driven" fashion. There is a real discovery inside that pitch, and it is not the one the summit's name advertises. Fashion's gravest environmental sin is making too many clothes; its gravest financial sin is the same thing. The lever that reduces the one reduces the other, which is why the prettiest-sounding idea in the industry is quietly the most profitable one.

The paradox at the heart of the industry

Consider the arithmetic of waste. The global industry turns out somewhere between 100 and 150 billion garments a year and is responsible for roughly a tenth of global carbon emissions. Fast-fashion retailers routinely overproduce by 30–40%, knowing surplus will go unsold; what does not sell at a discount is landfilled or, in darker reports, incinerated. Each of those outcomes is also a charge against the income statement — markdowns, write-offs, discounting, and the capital tied up in stock that earns nothing while it sits.

This is why the AI part of the sustainability story is not greenwash. A model that forecasts demand to the SKU, store and week attacks the surplus at its source, and does so in the direction that profits point. More than a third of fashion executives now say they use generative AI in daily operations, and rank it their biggest opportunity; the reason is not a sudden collective passion for the climate but that the tool pays for itself. Business as usual produces clothes nobody wants; the AI fix produces slightly fewer, sells more of them at full price, and frees cash.

The proof sits in a quarterly report

No company illustrates the point better than H&M, the Swedish mass-market retailer and one of the world's largest, which has spent years known less for sustainability than for the opposite: mountains of unsold stock. Its latest results show the new shape of the business. In the six months to the end of May, operating profit before one-off costs rose 14% to SEK 8.1bn; in the second quarter the operating margin reached 12.0%, up from 10.4% a year earlier, and its rolling twelve-month figure climbed from 6.5% to 8.5%. Gross margin improved by 120 basis points to 56.6%, on a higher share of full-price sales and fewer markdowns, while stock-in-trade fell to 15.8% of sales from 16.6%. Revenue, meanwhile, barely moved, falling 1% in local currencies over the half-year.

That configuration — fatter margins, leaner shelves, flat sales — is exactly the shape a business takes when it stops betting on demand and starts predicting it. H&M cut its store count by 3% and is buying closer to need. In the old "push" model, a retailer locked in orders months ahead and hoped; now the company buys against forecasts and accepts that it will hold less. The gain shows up as margin and cash today; the cost is deferred to the moments the forecast is wrong.

And here the honest tension appears. H&M's operating profit missed analysts' expectations precisely because the leaner inventory and smaller store base meant it could not "fully meet demand" in some lines; its chief executive, Daniel Ervér, conceded sales came in "somewhat lower than planned." Lean inventory is a reallocation of risk and reward. When demand is strong, cautious stock forfeits sales; when it is soft, it preserves margin. What looks like a green strategy and a margin strategy is, looked at closely, a bet that shoppers are more price-sensitive than the buy-inventory school believes. So a shareholder should ask not whether the margin is higher — it is — but whether it survives the next upswing, or whether it is partly a transfer from lost volume.

Where the glamour misleads

The summit's framing encourages a different, more seductive reading: that the future of sustainable fashion lies in clever materials and textile-to-textile recycling. The fate of Renewcell suggests otherwise. The Swedish pioneer of recycled cellulosic fibre, a darling of the circular-fashion movement, collapsed in bankruptcy in 2024 after failing to secure financing, a reminder that recycled fibre competes with virgin material that is almost free and enjoys little premium from customers. The glamorous frontier stayed uneconomic; the unglamorous one — predicting demand so that less is made at all — is the one that pays.

For an investor, the distinction matters. The economic heart of "sustainable fashion AI" is not a new fabric; it is buying discipline, and the companies that genuinely reward shareholders are those, like H&M, embedding it in their operations and letting it show in margin, markdowns and working capital rather than in a sustainability report. The pure-play circular start-ups, by contrast, sell a promise whose unit economics have not yet closed. Fashion will be greener to the extent that predicting demand becomes cheaper than clearing the leftover — which is the one reform the market is happy to fund. The summit called it sustainability. The results desk calls it margin.