Qwen: The AI That Will Buy You Lunch (No Joke)
Chinese giant Alibaba has decided the best way to a user's heart is free food. The "free cards" promotion in the Qwen app has been extended to the end of February because of runaway demand. The chatbot no longer just writes code: it now pays for your lunch, milk or vegetables at Hema stores. It is an aggressive move in the fight for mobile traffic. While rivals compare model parameters, Alibaba is building AI into everyday life through its retail ecosystem. If you thought AI was only about generating text, look at how Qwen is turning into a wallet with a brain.
AI-processed from 36Kr (36氪); edited by Hamidun News
Picture this: you open an AI app not to draft a workout plan or write Python code, but to pay for a dozen eggs and a carton of milk at the nearest supermarket. It sounds like a techno-optimist's strange dream, but for users of Alibaba's Qwen model in China, it is now reality. The company officially confirmed that the buzz around its holiday "free cards" promotion was so strong that it had to be extended on short notice. The campaign, an attraction of unprecedented generosity, was originally set to end on 23 February; the deadline has now moved to the 28th.
The context matters more than the discount itself. China is currently in a fierce war over the title of the country's "main AI assistant." While Baidu bets on integrating its Ernie Bot into search, Alibaba is playing its trump card: a giant retail ecosystem. A coupon worth 25 yuan is not just a discount on coffee. Qwen now lets users pay for full meals, vegetables, groceries and even goods on Tmall. What's more, the entire Hema chain of high-tech supermarkets is now officially integrated with the Qwen app. It is a textbook example of a corporation shifting the fight from cloud computing to real-world consumption.
Why is this happening now? The holiday period in China is a time of peak spending and maximum audience attention. The cost of acquiring one active mobile app user is sky-high today. Instead of burning budgets on classic outdoor advertising or banners, Alibaba is literally buying loyalty through the stomach. If users get used to an AI that helps them save on staples, they are unlikely to delete the app when the promotion ends. It is the classic "O2O" (Online-to-Offline) strategy, pushed to the extreme with artificial intelligence.
The gap in approaches is interesting to watch. In the West, OpenAI and Google fight over whose model passes the bar or medical exam better. Meanwhile, in China, AI is learning to be useful in the checkout line. That does not mean Qwen is dumber than its Western counterparts: its language abilities are impressive. Alibaba simply understands that to go truly mass-market, AI has to move beyond the browser and the chatbot. It has to become part of the physical world, where people buy dumplings and milk tea.
Qwen's link to Hema's offline stores shows where the industry is heading. We are entering an era in which the AI assistant stops being merely a "smart conversation partner" and becomes a full operating layer between people and services. Where you once hailed a taxi or ordered food through separate buttons, Alibaba is now trying to do it all through a single Qwen interface. And if that means handing out millions of free lunches, it will do so without hesitation. The only question is how long this subsidy party will last and what happens when the marketing budgets run out.
Bottom line: Alibaba is turning an AI model into a tool for everyday consumption, tying it to its commercial empire. While others teach AI to philosophize, Qwen is learning to buy milk. Will this become an industry standard, or remain an expensive traffic-acquisition promotion?
- Nov 2, 2026DeepSeek open-sources Super-agent prototype: p95 task completion time of 15.7 minutes vs. 29.5
- Nov 2, 2026TensorFold 1.2.0 boosts Qwen3.8-27B throughput on M3 Ultra by 44.2% at 16 streams
- Nov 1, 2026Adobe's EditAgent judge picks a usable edit 81.3% of the time, versus 74.8% for EditScore
- Nov 1, 202616 models tested on 7 search tools: Exa averaged 10.5 out of 14, SearXNG about 3
Test Frontier LLMs & Autonomous Agents in One Workspace
Run Claude Opus 5.5, GPT-6, Gemini 3.1, and DeepSeek side-by-side in Nomad AI (nomadnet.ai) — unified model routing, deep research, and multimodal workflows under one subscription.
The AI world, distilled — once a week
Seven stories that actually mattered, hand-picked. No noise, no reposts, no press releases.
Done! Check your inbox for a confirmation.