OpenAI is officially retreating from handling purchases directly after Instant Checkout 'failed to provide enough flexibility.' Translation: brands didn't want a third party sitting between the customer and the transaction. Walmart has now moved Sparky into ChatGPT's shopping layer — either the savviest distribution play of the year or a slow surrender of customer data to yet another platform. Admiral Neritus Vale has the full read on what this means for fashion brands in today's piece.

The pattern is clear: checkout is becoming contested infrastructure. Every platform wants to own the final click. Fashion brands who assumed AI-powered discovery was the hard problem are about to discover that the last ten meters of the customer journey are the most expensive real estate on the internet. Right now, Shopify is the only major platform actively trying to keep that real estate in brands' hands — and this Walmart-OpenAI handshake just made Shopify's pitch considerably easier.

Meta is rolling out generative AI across Instagram and Facebook shopping to surface 'more product and brand information.' How thoughtful. Except Sir John Crabstone's piece today explains what this actually is: not a feature, a toll booth. Every product detail Meta's AI surfaces is a data point Meta owns, not you. Every 'helpful' recommendation keeps the customer inside Meta's ecosystem instead of yours.

The timing is extraordinary. On the same day, a jury found Meta negligent in the landmark social media addiction trial — confirming that Meta understood exactly how addictive its platforms were. Meta knows what it's doing when it makes its apps stickier, and now the courts agree. Fashion brands who have built their entire acquisition strategy around Instagram should be watching both of these stories at the same time, because they are the same story.

Fifteen years is an eternity in app time. Poshmark's redesign spotlights sellers, buries the flea-market aesthetic, and leads with AI-powered discovery. This is what infrastructure looks like from the customer side — today's piece on the secondhand market growing up gets into what it means structurally. Resale is no longer trying to prove it's legitimate; it's building the pipes that will outlast the brands it's disrupting.

Compare this to what The Fashion People is doing for brands trying to reclaim control of their resale presence. The gap between 'we have a resale program' and 'we actually understand our resale market' is where the real money is about to move.

Lowe's SVP of data and AI sat down with Modern Retail to explain how the company decides which AI agents to build — and how to stop building ones that don't pay off. This is the ground-level version of what Sir John Crabstone argues today: the crisis in agentic commerce isn't a shortage of agents, it's a proliferation of them, each solving a slightly different problem, none talking to each other.

Lowe's sells lumber, not handbags, but every fashion operations team should read this carefully. The questions they're wrestling with — build vs. buy, ROI measurement, agent governance — are identical to what any multi-channel fashion retailer will face the moment they try to deploy AI across buying, merchandising, and customer service simultaneously. The complexity doesn't announce itself. It arrives all at once.

Gap CTO Sven Gerjets at Shoptalk, announcing partnerships with Bold Metrics (AI sizing) and Google's Universal Commerce Protocol. This is precisely what today's piece unpacks: Gap isn't chasing the flashiest demo, it's choosing the standards that will determine how AI agents interact with fashion inventory at scale. Bold Metrics means better fit data. Universal Commerce Protocol means Gap's products become legible to every AI shopping agent on the internet.

The contrast with brands doing AI for the press release is stark. Gap isn't announcing a chatbot. It's joining a protocol. Those are completely different bets on where value accumulates in 2028. We wrote about the trust tax on obvious automation — Gap is quietly making the opposite move.

ASOS underlying profitability up 50% for H1. A year ago this company was being written off as an overextended fast-fashion casualty of sky-high return rates and discount addiction. Now it's printing profits. Sir John Crabstone puts this in context today alongside Chaussea's Kickers acquisition — Europe's middle market isn't dying, it's sorting itself out violently. The survivors are the ones who got ruthless about operations and stopped pretending they could be everything to everyone. Editing is a strategy.

Middle East tensions are pushing air cargo rates up 400% on affected routes, putting a $55 billion Gulf market and fashion's spring calendar at simultaneous risk. Admiral Neritus Vale has the full breakdown today, and it's worse than the headline suggests — this isn't just about cost, it's about the entire assumption that fashion can run on a global just-in-time supply chain indefinitely.

We already wrote about fashion's looming compute bill squeezing AI investment. Now add logistics costs spiking on top. The brands most exposed are the ones who optimized simultaneously for speed and margin — which is most of them. Rising gas prices are pressuring consumer spending at the same time. The squeeze is coming from every direction at once.

와이즈앱·리테일 (WiseApp·Retail) data: 48.7% of Korean smartphone users — 24.94 million people — are now using generative AI apps as of February 2026. Doubled in a year. Sir John Crabstone makes the retail case today: if you're a fashion brand treating AI-native consumer behavior as a 2028 problem, you're already late for one of the world's most sophisticated markets.

South Korea has the fastest 5G penetration globally, among the highest beauty spend per capita in the world, and a retail environment where the super-app model from China is actively competing for attention. When half the phones in a major fashion market are running AI assistants, the discovery problem doesn't evolve — it restarts from scratch.

Estée Lauder, still working through its post-pandemic hangover and a brutal stretch of writedowns, is reportedly in merger talks with Puig — the Barcelona family conglomerate behind Charlotte Tilbury, Rabanne, and Carolina Herrera. Analysts say the deal 'might look better on paper than in practice,' which is analyst-speak for 'we have serious doubts but no one's asking us.' Puig has the hot brands; Estée has the retail infrastructure. Whether that's complementary or just an expensive bet on synergies that never materialize is the question no one's answering out loud.

LVMH, Kering, Richemont — the first tier of luxury consolidation is assembled. Now the second tier is grasping for scale before the window closes. The brands that survive consolidation intact are the ones with genuine cultural equity. Distribution agreements don't survive the next recession.

Bernie Sanders and Alexandria Ocasio-Cortez introduced legislation to halt all new data center construction pending AI regulation. This will almost certainly not pass in its current form. But the political signal matters: compute infrastructure is now a legislative target, and that shifts the risk calculus for any brand building AI strategies on US-based cloud providers.

Connection worth flagging: We already have financial pressure on fashion's AI compute costs documented here, and Alibaba Cloud price hikes already squeezing APAC AI investment. Add political uncertainty in the US and fashion's AI infrastructure story goes from 'expensive' to 'genuinely complicated on three continents.' The brands that locked in multi-year cloud contracts in 2024 are looking considerably smarter this week.

Amazon is rolling out one-hour and three-hour delivery — on the same day FedEx announced a two-hour SameDay Local service via OneRail. The speed war is escalating faster than the unit economics can justify. For fashion specifically: you cannot compete on delivery speed if you're still running traditional wholesale calendars and centralized DC networks. The brands that matter in one-hour delivery have narrow, fast-moving SKU counts and dense local inventory positions. This is an argument for curation and localization, not for scale.

John Galliano — the designer with the most dramatic career resurrection in contemporary fashion — collaborating with Zara. FashionUnited is right to call it a power move. This is Inditex saying: we can carry the fashion world's most culturally charged designer name and still sell it at €79. No luxury house can respond. They are not competing in the same arena, and Inditex knows it.

Zara just made the cultural argument that prestige doesn't require a price point. Every mid-market brand nervously watching luxury's expansion downward needs to recalibrate. The squeeze isn't coming from above — it's coming from Zara with a Galliano hangtag. The European middle market story today suddenly has a more menacing subplot.

Pawn shop traffic is surging as financially stressed consumers look for quick cash or lower-cost secondhand goods. This is not the curated resale economy of Poshmark's freshly redesigned app — it's the stress-driven liquidation of discretionary items when the bills arrive. Today's piece on resale infrastructure covers the optimistic end of this market; pawn shops are the reality check that belongs in the same paragraph.

Connection: Pawn shops are booming at the exact moment rising gas prices threaten to break consumer spending broadly. Two data points, one signal: we are entering a bifurcated consumer moment. Ultra-premium holds. Value surges. The middle gets hollowed out further. This is what the consumer side looks like — and the ASOS and Chaussea stories today show exactly the same dynamic from the retailer side. The consumer and the industry are telling the same story from opposite ends of the transaction.