Securities Times is running what amounts to a strategic autopsy on Alibaba's Taotian group — Taobao and Tmall bundled as a "dual engine" — while the 380-billion-yuan AI bet gets made on top of a business that is visibly stalling. The piece's framing is precise: the damage is in growth rates, not absolute scale. JD is poaching merchants with logistics guarantees Alibaba cannot match on speed. Pinduoduo is grinding down on price. Douyin is eating the discovery layer that used to be Taobao's monopoly. Securities Times adds a personnel dimension we hadn't seen: JD is specifically recruiting Taobao's supply chain talent, which means the institutional knowledge embedded in Alibaba's merchant relationships is walking out the door on a human schedule.

We covered the structural version in Alibaba Pays Retail for Growth Taobao Used to Collect Free and the capex question in Alibaba Is Paying 380 Billion Yuan for a Moat Its Rivals Do Not Need. The capex answer is now clearer: the 380 billion goes into model training, inference, and the Qwen stack. None of it directly fixes merchant retention or rebuilds the discovery edge. Hillhouse and Jinglin's 13F positions already showed where smart money thinks this ends. The parcel beats the cloud.

At Capital Markets Day, Kering detailed how its L'Oréal partnership will scale its beauty business. The logic is impeccable and slightly uncomfortable: de Meo is rebuilding Kering like a car company, which means outsourcing components that are not core competencies. L'Oréal gets licensing fees and access to Kering's brand equity. Kering gets beauty revenue without building a beauty operation from scratch while simultaneously fixing Gucci, stabilising Balenciaga, and diagnosing McQueen per the capital markets triage from last week. LVMH builds its own beauty pipeline — Parfums Christian Dior, Guerlain, Givenchy Beauty. Kering is renting the capability. That gap compounds over fifteen years in ways that are invisible at the quarterly level and decisive at the decade level.

Two separate companies, same anxiety: every platform that missed TikTok Shop's 2022–2024 window is now bolting commerce onto whatever discovery surface it already owns. Pinterest has 300 million users who arrive specifically to save things they intend to buy — the question was always whether the transaction would close in-session. Live shopping is one answer. Shopify's Agentic Storefronts are another. John Lewis's TikTok Shop pilot is a third — and today we're looking more closely at exactly what John Lewis bought with that £800m. All of them are trying to own the moment before the customer opens a new tab.

The BoF opinion piece presumably means AI agents are disintermediating traditional e-commerce funnels — which is correct, and genuinely important — but calling online shopping a "victim" of AI is like calling email a "victim" of Slack. The medium shifts; the behavior persists. Honor's MagicOS agent shops before Taobao sees the query. David's Bridal routes checkout through ChatGPT. Shopping is still happening. What is getting disrupted is the discovery tax that platforms have been collecting — the toll between intent and transaction. That is not shopping dying. That is the middlemen losing their rent.

The gap between what AI can do online and what exists in a physical store is still enormous, and whoever closes it first controls the in-store data layer. Dollar General's 12,000-store audio network — no screens, just influence — is one approach. VF's item-level RFID layer with Nedap is the inventory side of the same bet. The brands building AI infrastructure in stores right now are also building the training data for the next generation of models, a compounding advantage that pure e-commerce players cannot close by adding more GPUs.

FashionUnited asks who is actually shopping in luxury boutiques. We answered the structural version three days ago: the queue has become the product, and brands are paying Place Vendôme rent to stage a line that may never convert. The FashionUnited piece adds the granularity: it is increasingly tourists and one-time gift buyers rather than repeat core customers. Which means the flagship is doing brand-awareness work at prices built to justify sales-floor work. The lease portfolios of every major luxury house will be restructured as that math becomes impossible to ignore. Moncler's 12% quarter — covered today — is one data point on what decoupling from the flagship dependency actually looks like in the numbers.

Innovative Eyewear posted record preliminary Q1 2026 sales. The pattern we flagged in A Frame You Forget Beat A 750-Gram Headset is holding: smart eyewear wins when it looks like eyewear. EssilorLuxottica's Ray-Ban Meta units tripled to seven million. Innovative Eyewear's Lucyd line is growing the same lane. The headset makers are still learning that the distribution channel for wearable AI is the optician's appointment book, not the consumer electronics shelf.

Southeast Asian e-commerce is projected at 16% growth this year as video commerce and AI reshape the shopping surface. Today we publish on Vietnam's $110B infrastructure bet that skips the wholesale layer entirely — the regional story is that every market from Jakarta to Ho Chi Minh City is constructing logistics and payments infrastructure that bypasses the distribution middlemen Western markets took five decades to slowly disintermediate. Video is the discovery layer. TikTok Shop is the transaction layer. Both are newer than the category management systems most Western brands still run on. That catch-up dynamic does not last forever. While it does, the structural advantage belongs to whoever builds the infrastructure.

Shark Beauty and Lookfantastic are adding live social commerce to Topshop's AI-driven virtual catwalk. This is the most direct Western attempt yet to replicate Chinese live commerce mechanics in a fashion context — Pinduoduo's interactive drama routes plot choices to product pages; this routes catwalk looks to purchase buttons. The format difference matters enormously: Chinese live commerce works because of a decade of consumer training, established trust signals, and a viewer base that long ago accepted the transaction as part of the entertainment. Topshop is starting from zero on all three counts. The novelty carries the first event. The second event is where we learn if there is a product here or just a press release.

Uniqlo is launching with Danish designer Cecilie Bahnsen — voluminous tulle, extreme femininity, the categorical opposite of LifeWear utility. This is deliberate. Bahnsen's customer cannot buy Uniqlo without some cognitive dissonance; Uniqlo's customer gets an introduction to a designer they could not otherwise afford. We wrote about Uniqlo's Manhattan neighborhood cluster strategy importing Starbucks visit-frequency math. This collab is the reason to make the visit — and then the reason to come back the following week for the Supima cotton T-shirts that are still on the rack after the collaboration pieces are long gone.

Reale Actives cleared inventory on launch day, collected its mixed reviews, and is now in the hard part: converting a TikTok audience into a repeat customer base. The CEO saying they "prepared for a mixed response" is genuinely a good sign — it means they understand that the influencer funnel is a launch mechanism, not a retention mechanism. Wonderskin's second act — covered today — is the best case study in this zip code right now. The peel-off lip stain went viral and then the brand had to figure out what else it was. Alix Earle has the audience. She needs the portfolio depth.

The connection worth flagging: Wonderskin, Reale Actives, and I.AM.GIA are all solving the same problem with different bets — I.AM.GIA went concentration (one product, 300,000 units, literally the founder's house as collateral). Diversification is the other path. Neither is obviously safer. Both require the founder to make the call before the data arrives.

Dr. Joyce Park's Kerativ enters a category that Sephora is actively doubling down on for margin — covered today in Sephora Doubled K-Beauty Hair To Find Margin. Dermatologist-founded brands have proven durable in skincare (Dr. Dennis Gross, Dr. Muneeb Shah's Remedy Skin). Hair is harder: the claim language is regulated differently, the ingredient education curve is steeper, and the K-beauty players spent three years building the category narrative at Sephora before Western consumers caught on. Credentials are the entry ticket. Distribution is the actual barrier. The question for Kerativ is whether the dermatologist positioning translates into purchase intent in a category where most buyers are currently reaching for something they read about in a Korean Reddit thread.

This week's moves include Debenhams Group, JD Sports, Lululemon, and Vuori. Debenhams is the one to watch: we covered their decision to place marketplace technology and AI under a single CTO in April, and continued senior-leadership buildout in the same direction suggests the organizational structure is being constructed ahead of the product strategy — which is either genuine conviction or elaborate cargo cult, and you usually don't know which for two years. Lululemon and Vuori appearing in the same weekly roundup is the athleisure succession story in miniature: one brand defending premium positioning under pressure, one challenger growing into the same lane with a different aesthetic and a cleaner balance sheet.