Alibaba's Qwen Now Pays for Your Groceries

Alibaba is raising the stakes in the fight for user loyalty. The Qwen app has extended its "free orders" program to the entire Tmall Supermarket catalog. Users can now order groceries, alcohol and household goods with a single phrase in the chat, and the company pays the full bill. This is not just a marketing promotion. It is an attempt by Alibaba to make its language model the central hub of its whole ecosystem. While competitors train models on abstract data, Alibaba is teaching millions of people that AI is not only text but also real food delivered to their door. The strategy is aggressive and expensive, but amid fierce competition with Baidu and Tencent it looks like the only way to hold on to an audience.

AI-processed from 36Kr (36氪); edited by Hamidun News
Alibaba's Qwen Now Pays for Your Groceries
Source: 36Kr (36氪). Collage: Hamidun News.
Listen to article

Remember when we were amazed that a neural network could write a coherent email or draw a cat? Forget it. While Silicon Valley holds philosophical debates about whether AI will replace programmers, the Chinese tech giant Alibaba has decided to approach the question in the most down-to-earth way possible. Its flagship Qwen app now, quite literally, feeds its users.

The company has announced an expansion of its "free orders" program, integrating the chatbot with the giant Tmall Supermarket marketplace. Now you only need to say one phrase into the microphone, and an hour later groceries, household cleaning products or even fresh fruit will arrive at your door, with not a single yuan to pay.

To understand why Alibaba is burning millions on free dumplings and rice, you need to look at the context of the Chinese market. The battle for leadership in large language models (LLMs) has entered a phase of total attrition warfare there. Baidu with its Ernie Bot and Tencent with its cloud solutions are breathing down its neck. In these conditions, a merely "smart chatbot" is no longer enough. Users need a practical reason to open the app every day.

Alibaba is betting on seamless integration of its ecosystem. Previously, buying groceries meant going to a website, searching for a product and building a cart. Now Qwen takes on the role of a concierge who not only understands context but also has signing authority over your receipt.

Technically, this looks like an ideal test for "agents": AI models that can take actions in the real world. Alibaba is testing not just language understanding but the reliability of transaction execution. When you say "order me the same chips as last time, and add milk," the model has to match your order history, the stock available in the Tmall Supermarket warehouse and current discounts. That is harder than generating poems, because the price of a mistake here is real logistics costs and an unhappy customer.

Extending the program to categories such as alcohol, fresh food and household goods suggests Alibaba is confident in the accuracy of its algorithms.

For the industry, this is an important signal. We are watching the end of the era of "pure AI" and a shift toward AI services. Other players will most likely soon adopt similar mechanics, turning their models into operating systems for daily life. Alibaba understands that data on which products you buy through a chatbot is worth far more than a couple of free packs of rice.

This is a battle for habit. Whoever becomes your main interface to the physical world will ultimately take everything. While Western companies fear the legal consequences of every step, Chinese giants are simply buying audience loyalty, in the most literal sense.

Bottom line: Alibaba is turning AI from a toy into a payment tool. Can Qwen become a full replacement for traditional shopping, or is this just a very expensive way to pump up active-user metrics?

Story
ZK
Zhemal Khamidun — CPO at Alpina Digital, product manager of AlpinaGPT. Building AI products since 2019. Hamidun News is a pet project that grew into a production aggregator with 6,000+ readers.
✦ Unified AI Workspace · Nomad

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.

Try Nomad AI →Launch Nomad Agent
What do you think?
Loading comments…