A tariff wall on one side, virtual fitting rooms on the other—who’s actually making fashion easier to buy? If you’re new to this story, here’s where we are. Fashion’s first DTC wave is getting repriced: Everlane’s reported $80 million sale to Shein, Outdoor Voices’ store closures and acquisition, Parade’s sale, and Allbirds’ reset have all shown the limits of online-only growth. Cuyana and Faherty are leaning on repeat customers and sell-through, with stores and wholesale in the mix, rather than pure distribution mythology. This is Fashion Business Daily. Today, the border scramble is hitting margins. We’re also into the AI fitting-room pileup—and a sample sale that got very, very real. Danny Parisi, writing in Glossy:
In the last few weeks, tensions between the two countries over the Trump administration’s aggressive tariffs on Canadian goods boiled over into a full-on trade war. Canadian negotiators rejected a deal they called “unfair” and unserious. In response, the U.S. enacted a 50% tariff on nearly $30 billion worth of Canadian goods exported to the U.S., targeting a wide array of exports including clothing and apparel. Canada responded by promising “ dollar for dollar” counter tariffs on U.S. goods.
A 50% tariff on nearly $30 billion of Canadian goods hits margins hard. Outerwear and men’s formalwear are specifically exposed, and Canada’s $2.8 billion in apparel exports to the U.S. now comes with a very different pricing equation. Ssense and Duer are already rerouting fulfillment. That’s supply-chain surgery — and somebody ends up paying for it through a thinner assortment, a higher ticket, or both. The issue is shipping from Canada, regardless of where the garment was made. A Canadian label sourcing overseas still gets caught at the border. Management teams need to spell out category exposure and gross-margin impact instead of gesturing at a less-integrated market. Canada Goose and Arc’teryx built real equity around making cold-weather gear in Canada. A tariff wall puts a price on that origin story — please don’t solve it by squeezing the people making the parkas. This one's from The Star:
KUALA LUMPUR: New York-based fashion technology startup Lekondo, founded by Malaysian entrepreneur Yeng Tan, has raised about US$3mil from investors including Verdict Capital, Andreessen Horowitz (a16z) SR and DG Daiwa Ventures. In a statement, the company said the funding would support the development and growth of its fashion platform, which allows users to post daily outfits, build digital closets and discover how people dress around the world.
Lekondo has hit 100,000 users in 50-plus countries since January by asking people to post what they actually wore. I like that more than another app pretending a saved product page is a personality. The company has raised about $3 million from Verdict Capital, a16z SR and DG Daiwa Ventures. That user growth is company-reported—not proof of revenue or retention, or that brands will pay for this data. Digital closets can become a very expensive mirror. But if Lekondo can surface how people style one jacket ten different ways, that’s richer product intelligence than watching somebody panic-buy a trench at 1 a.m. Yeng Tan’s pitch is that worn outfits reveal taste better than clicks and purchases. Fine. Bring cohort retention and evidence that discovery leads to conversion, then we can price the signal. Genlook writes:
GRENOBLE, FR — September 2, 2026 — Genlook released a new version of its AI try-on model today. It generates a photorealistic image of any garment in a store’s catalog on any shopper’s body under 10 seconds, from a single photo they upload once. The new model supports every type of fashion garment, from dresses and outerwear to swimwear, footwear and accessories, including categories earlier versions could not render reliably.
Genlook says it’s live in more than 600 fashion stores, with try-on users adding to cart at three times the rate. It’s a company-reported claim, not an independent study—and add-to-cart is the easy part. Right. Show me the returns data. If somebody can flick through a whole catalog in ten seconds and still orders two sizes, Genlook has made the shopping dopamine faster and solved exactly nothing downstream. One uploaded photo, any garment, including swimwear and footwear—that’s a much broader product promise than the earlier try-on tools. Before I buy it, show me a control group. Then show conversion past the cart and whether return rates hold at scale. And if the fit image is wrong, the warehouse, the driver, and the person processing that return eat the consequences. A photorealistic picture is cute, but what matters is whether the jacket fits when it lands. Here's EXA:
Methods and systems are disclosed for using generative machine learning models to generate fashion items for avatars. The methods and systems present a graphical user interface (GUI) comprising icons representing different types of avatar fashion items. The methods and systems receive input that selects an individual icon corresponding to an individual avatar fashion item.
Snap got U.S. Patent 12,725,332 on Monday for generating avatar clothes from prompts and textures. Cool—until every avatar is wearing prompt-engineered beige nylon with the personality of a loading screen. Keep this separate from the Genlook rollout we just covered: Snap has protected an interface and a generative-ML method, filed in April 2024. For now, a granted patent is IP positioning; it tells us nothing about shopper conversion or return rates. We’ve got three avatar and try-on bets in two days, and everybody wants fashion to be a texture menu. Brands still need somebody with taste deciding whether the jacket is worth making in the first place. This one's from Modern Retail:
A four-block line. A two-and-a-half hour wait time. And hundreds of towels, candles and cashmere sweaters. These were the key elements of Quince’s first-ever sample sale, held Aug. 14-15 in its hometown of San Francisco. Quince, which officially launched in 2020, is a digitally native lifestyle brand known for its $50 cashmere sweaters. Now, after surpassing $2 billion in sales, the brand is increasingly testing out physical retail.
Quince says it has crossed $2 billion in sales, then its first San Francisco sample sale cleared out six hours early. Six hundred transactions at three items apiece is a real offline signal—especially with a four-block line. People waited two and a half hours for photo-shoot cashmere, towels, and coats. The $50 sweater has escaped the browser tab. For the DTC omnichannel reset, this is useful proof. Pop-ups gave Quince a test run, and this sale suggests shoppers will show up in person. But it was finite sample inventory, so I’d want the next store test measured against normal-price sell-through. And Quince’s whole pitch is same materials for less money. Now let’s see whether that $2 billion machine can keep quality up as it expands categories, without squeezing the factories that make the cashmere possible. If you’re enjoying Fashion Business Daily, please subscribe or leave us a review wherever you’re listening. Reviews help other people find the show, and we’re grateful you’re here.
Links to every story are in the show notes, so take a look at the pieces that caught your attention and explore the details at your own pace. That’s Fashion Business Daily for today. This is a Lantern Podcast.