
Happy Sunday! DoorDash is turning local Shopify inventory into a native marketplace channel with one-hour delivery. Retailers gain rented demand and logistics, but the margin test still arrives after commissions, picking work, and refunds.
In today's MarketMaze:
News
1️⃣ DoorDash turns stores into channels
2️⃣ Cohen turns 9.8% into pressure
3️⃣ Europe opens Google's search vault
Insights
4️⃣ AI capex outruns every precedent
5️⃣ AI shopping splits in two
6️⃣ Shein's margin meets reality
LET’S ENTER THE MAZE!
1️⃣ News

The Maze: DoorDash is now a native Shopify sales channel for U.S. physical retailers, syncing local inventory into delivery orders inside one workflow.
Merchants can publish selected products without separate onboarding or manual catalog uploads, then have a Dasher collect packed orders from the store.
The app syncs products, inventory and orders, with delivery in as little as one hour and access to DoorDash's company-reported 42 million monthly users.
Installation is free, but DoorDash commissions vary; merchants still carry picking labor, stock accuracy, refunds and the risk that marketplace sales do not clear the margin hurdle.
Why it matters: Local retailers can rent demand and delivery instead of building them. The real test is not speed, but incremental profit after fees and fulfillment work.
2️⃣ News

The Maze: GameStop converted its eBay derivatives into a 9.8% voting stake. Cohen has real shareholder leverage now, but still no control or agreed deal.
GameStop owns 43.4 million eBay shares after deploying more than $4.3 billion from working-capital cash.
Cohen wants 1,600 US stores to authenticate collectibles, handle shipping and host live commerce on eBay's marketplace rails.
eBay rejected the $125-a-share proposal over financing, leverage, valuation and execution concerns.
Why it matters: A 9.8% stake turns takeover rhetoric into shareholder pressure. The commerce thesis has logic; the financing and $2 billion savings promise remain unproven.
3️⃣ News

The Maze: The EU will make Google share anonymised search feedback with eligible rivals. It is opening the learning loop, not the algorithm.
The binding rules cover ranking, query, click and view data for search engines and search-enabled AI chatbots from January 2027.
Access is ringfenced: identifiers are removed, rare queries suppressed, recipients vetted and independent audits required.
Google can recover incremental costs. Data is delayed at least seven days, limited to five years and cannot train general-purpose AI or clone Google results, as the lead report explains.
Why it matters: Search quality compounds through feedback. Europe is testing whether regulated access can create credible rivals without making private searches public.
4️⃣ Insight

The Maze: AI capex reached 4.46× its 2023 trough in three years. Canal mania needed five years to peak at 4.09×. Speed is now the risk, while the revenue clock lags.
The BIS series puts railway mania at a 2.65× peak, while dotcom investment never reached 2×.
The AI path ends in 2026; it is not a forecast, and each historical line uses a different investment base normalised to its own trough.
A winner-take-most race can make every firm invest rationally while the sector still builds too much capacity too early.
Why it matters: AI can transform the economy and still punish badly timed capital. The next moat is not more compute. It is utilisation that pays for it.
5️⃣ Insight

The Maze: AI research is spreading globally, but checkout is not: China is building one commerce stack while Western shoppers still cross between assistants and stores.
China leads direct AI purchase at 23%, versus 15% in the US and 12% in the UK.
Product comparison is much closer at 35%, 29%, and 30%, so the real divide appears when AI must act, not merely advise.
Alibaba gives Qwen 4 billion listings, while ChatGPT must connect merchants across a fragmented Western stack.
Why it matters: Brands need two playbooks: win inside China's integrated ecosystems, and make product data discoverable across the West's messier AI-to-retailer handoffs.
6️⃣ Insight

The Maze: Shein may enter Hong Kong at $40B-$50B, half its 2022 peak. Revenue stayed enormous. Yet the profit expected to justify the old price did not.
The source comparison puts 2024 revenue at $38B versus a $45B target, a miss of roughly 16%.
Net profit landed near $1B versus a $4.8B target, while the Singapore filing is cited at $1.29B.
The EU's EUR3 low-value parcel fee now pressures a model that gets about one third of revenue from Europe.
Why it matters: Shein proved demand at global scale. Its IPO must prove the margin survives when regulators tax the cross-border shortcut that built it.
🗞️ Quick hits
Everything else you should know
💳 Agentic commerce gets funded
Airwallex raised US$320 million to expand payments infrastructure and autonomous finance, giving agentic commerce another well-capitalised transaction layer.
🏪 Marketplace power meets politics
🤖 AI moves into operations
Myntra expanded AI across seller onboarding and catalogue production, cutting turnaround from one day to four hours and producing 400-600 automated videos daily.
Evri is deploying Microsoft AI and five planned super-agents for routing, routine work, and parcel communications, pushing automation into delivery economics.
📦 Distribution stretches farther
Amazon is opening more Business capabilities to third-party sellers ahead of its October buyer conference, using marketplace inventory to deepen procurement demand.
India Post and Flipkart signed a logistics agreement, linking marketplace demand to the national postal network and widening seller reach beyond the largest cities.
THAT’S IT FOR TODAY!
You’re the reason our team spends hundreds of hours every week researching and writing this email.
What do you think of this issue?
See you next time in the maze!
MarketMaze team


