Search
left arrowBack
Eugene Levitin

Eugene Levitin

March 5, 2026 ・ Agentic Commerce

China Is Already Living in the Agentic Commerce Future

China Is Already Living in the Agentic Commerce Future

A VC from Hong Kong messaged me last week about Chinese vendors who want to sell to American AI shopping agents.

I didn't know much about agentic commerce in China. So I went digging.

Alipay processed 120 million AI-agent transactions in a single week in February 2026. Not browsing sessions. Not chatbot conversations. Completed purchases — ordered through a chatbot, paid without leaving the conversation.

In the US that same week, Google was announcing its protocol partners. OpenAI was onboarding its first Etsy merchants for ChatGPT checkout. We were launching waitlists. China was processing 120 million orders.

That number changed how I think about the agentic commerce timeline. Not whether it's coming — but how far behind the US already is.

The Scale of What China Built

The numbers across China's tech giants are hard to fully absorb.

Alibaba's Qwen chatbot hit 100 million monthly active users in its first two months after launching in November 2025. Users can compare product recommendations from Taobao, book flights on Fliggy, and pay via Alipay — all without leaving the chatbot. During Chinese New Year, Alibaba ran a 3 billion yuan ($431 million) campaign subsidizing purchases through Qwen. Over 10 million bubble tea orders were placed within 9 hours of the campaign launch — enough to crash the system and temporarily suspend coupon distribution.

ByteDance's Doubao agent reached 159 million monthly active users by October 2025. It can autonomously open four shopping platforms simultaneously — JD.com, Taobao, Pinduoduo, and Douyin Mall — scan screens, identify products, compare final prices including coupons, and present the cheapest option for one-click purchase. The whole process takes under 30 seconds.

Tencent integrated its Yuanbao AI chatbot directly into WeChat, giving 1.41 billion users access without downloading a separate app. Their Chinese New Year red-envelope campaign — 1 billion yuan ($144 million) — pushed Yuanbao to number one in Apple's App Store.

Baidu's ERNIE Bot has 200 million users. During Double 11, 83% of livestreamers used Baidu's digital human technology, driving a 119% year-over-year increase in participating livestreams and a 91% boost in gross merchandise value.

China's generative AI user base reached 515 million by June 2025 — 266 million new users in just six months. Market penetration: 36.5%.

The Super-App Architecture Advantage

Here's the part that matters most for anyone in ecommerce outside China.

China doesn't need a Universal Commerce Protocol because Alibaba already IS the protocol. When your AI chatbot (Qwen), your marketplace (Taobao), and your payment system (Alipay) are the same company, you skip years of interoperability debates.

In the US, the architecture looks fundamentally different. UCP connects Google to Shopify to Visa to Target — separate companies, each with their own systems, agreeing on a common standard. ACP connects OpenAI to Stripe to Etsy to the merchant's checkout. Every connection is a negotiation. Every integration is a project.

In China, Luckin Coffee's AI assistant handles the entire purchase flow — a user says "order my usual" in natural language, the AI selects the drink, confirms identity, processes payment through Alipay, and the order is placed. No checkout page. No manual confirmation. No leaving the conversation. According to Finextra, this was the first implementation where an AI agent handles both ordering and payment within a single conversational interface.

The US equivalent would require ChatGPT to talk to Shopify's API, authenticate through Stripe, verify the user's identity, check inventory, calculate shipping, and process payment — across at least four separate companies. We're building that. But China shipped it.

$647 Million in a Single Holiday Period

The spending numbers tell you how seriously China's tech giants take this.

During Chinese New Year 2026, Alibaba spent 3 billion yuan ($431 million), Tencent spent 1 billion yuan ($144 million), Baidu spent 500 million yuan ($72 million), and ByteDance partnered with the CCTV Spring Festival Gala for an undisclosed amount. Total disclosed spending: approximately 4.5 billion yuan — $647 million in a single holiday period.

These weren't marketing budgets. According to Inside Retail Asia, industry analysts call this a "key strategic juncture" that mirrors "the country's fierce battle for digital payment gateways over a decade ago" — when WeChat Pay and Alipay fought for mobile payment dominance. That fight created the infrastructure that makes today's AI commerce possible. This one is creating the habits.

The monthly burn rate across all tech titans: approximately $42 million per month on AI shopping tools, according to China Daily. They're subsidizing a behavioral shift — getting hundreds of millions of people comfortable buying things through conversations with AI.

Why the US Can't Just Copy This

The structural differences between China and the US aren't about technology or investment. They're about architecture.

China has super-apps. The US has apps. WeChat alone combines messaging, payments, mini-programs, and commerce for 1.41 billion users. WeChat Mini Programs handled 2.7 trillion RMB in transactions in 2023, with over 3.5 million active programs. Alipay and WeChat Pay account for over 90% of mobile transactions in China. There's no US equivalent — not even close.

China has integrated ecosystems. The US has competing standards. The US now has UCP from Google and Shopify, ACP from OpenAI and Stripe, MCP from Anthropic, A2A from Google, Visa's Trusted Agent Protocol, and Mastercard's Agent Pay. Six protocols, none fully interoperable. China has companies that own the whole stack.

China has consumer habits built on mobile payments. Mobile payments made up 73.2% of total transactions in China in 2024. Chinese consumers are already comfortable paying through conversational interfaces. American consumers are still figuring out whether they trust Apple Pay at checkout.

As one CNBC analysis put it: "The race is being won not by the most advanced AI model, but by the most integrated ecosystem." Western companies face "more fragmented data and stricter privacy regulations," slowing cross-service integration.

What This Means for US Ecommerce

The implication isn't that China is "better" at this. It's that the two markets are solving fundamentally different problems.

China's problem was: "How do we make AI commerce work inside ecosystems we already control?" That problem is solved. 120 million transactions a week.

America's problem is: "How do we make AI commerce work across thousands of independent companies that don't trust each other?" That problem is much harder. And the protocols being built — UCP, ACP — are genuinely impressive attempts at solving it.

But the timeline matters. Chinese platforms are perfecting buy-through-conversation on 515 million domestic users right now. Once it works at home, the export playbook writes itself. Alibaba is targeting nationwide rollout before 2026 Singles' Day. Baidu already launched its MeDo digital human agents in Brazil. ByteDance is developing a second-generation AI smartphone with ZTE, with Doubao pre-installed and agentic commerce coupons for first-time shoppers.

Temu has already caught up with Amazon in global cross-border ecommerce share. The US ended the de minimis exemption for shipments from China and Hong Kong in May 2025, applying duties to millions of low-value parcels. But tariffs don't stop the technology. Chinese sellers are already using AI tools for English-language product listings and customer service targeting Western markets.

The VC who messaged me wasn't asking a theoretical question. His portfolio companies are already planning for it.

The Uncomfortable Comparison

| Metric | China (Feb 2026) | US (Feb 2026) | |--------|------------------|---------------| | Weekly agentic transactions | 120 million+ (Alipay alone) | No comparable metric exists | | AI shopping user base | 515 million GenAI users | 39% use AI for product discovery | | Infrastructure spending | $647M+ in one holiday period | Protocol development stage | | End-to-end agent commerce | Live — order, pay, deliver in one flow | Experimental (ChatGPT checkout, Perplexity shopping) | | Time to complete purchase | Under 30 seconds via chatbot | Still requires multiple apps |

The US has advantages China doesn't: open protocols that could create a more competitive market, stronger privacy protections, and a more diverse merchant ecosystem. But on raw execution speed, it's not close.

What US Sellers Should Watch

The cross-border threat is real but not immediate. Chinese agentic commerce is focused domestically right now. But once Alibaba and ByteDance perfect the model, export follows. US sellers should watch for AI-native competitors entering their categories through cross-border platforms.

Protocol adoption matters more than ever. If US commerce stays fragmented while Chinese ecosystems ship integrated solutions, the gap widens. Getting your store onto UCP and ACP isn't just about ChatGPT visibility — it's about building the connective tissue the US market needs to compete.

Product data is your moat regardless of geography. AI agents in every market need the same thing: structured, accurate, real-time product data. Whether the agent runs on Qwen or ChatGPT, it reads schema.org markup. Get your structured data right first.

Agentic commerce starts with commodities. In both China and the US, AI purchasing works for replenishment and low-stakes items. If you sell commodity products, this is your timeline. If you sell considered purchases, you have more time — but not as much as you think.

FAQ

How far ahead is China in agentic commerce?

China is significantly ahead in transaction volume and consumer adoption. Alipay processed 120 million AI-agent transactions in a single week in February 2026 — the US has no comparable metric. China's generative AI user base reached 515 million by mid-2025, and super-app architecture means AI agents can complete end-to-end purchases in under 30 seconds. The US is still in the protocol-building and pilot stage.

Why can't the US replicate China's super-app approach?

The US lacks the integrated super-app ecosystems that make China's agentic commerce possible. Chinese platforms like Alibaba own the chatbot, marketplace, and payment system in one company. The US market is fragmented across separate companies (Google, Shopify, Stripe, OpenAI) that must agree on protocols to interoperate. Stricter privacy regulations and different consumer payment habits also slow integration.

Is Chinese agentic commerce coming to US markets?

Not yet directly, but the groundwork is being laid. Temu has caught up with Amazon in cross-border ecommerce share. Baidu launched AI commerce agents in Brazil. ByteDance is developing AI smartphones with pre-installed shopping agents. Chinese sellers are already using AI for English-language product listings and customer service. Once domestic AI commerce models are proven, cross-border expansion is the logical next step.

How much did Chinese tech giants spend on AI shopping in 2026?

During Chinese New Year 2026 alone, Chinese tech giants spent approximately $647 million — Alibaba ($431M), Tencent ($144M), and Baidu ($72M) — subsidizing AI shopping adoption. The monthly burn rate across all tech titans is approximately $42 million per month on AI shopping tools, according to China Daily. Analysts compare this to the decade-old battle for mobile payment dominance that created China's current payment infrastructure.

What should US ecommerce sellers do about the China gap?

Focus on what you can control: get your product data structured and machine-readable (schema.org JSON-LD, GTINs, real-time pricing), adopt UCP and ACP protocols as they become available, and submit products to Google Merchant Center. The structural advantages that make Chinese agentic commerce fast — super-app integration, unified payments — aren't replicable. But clean product data works for AI agents in every market.


Sources: Alipay AI Pay Milestone (Business Wire/Morningstar, Feb 2026), CNBC Chinese Tech Giants Report (Jan 2026), Caixin Global Alibaba Integration (Jan 2026), Inside Retail Asia Chinese New Year Coverage (Feb 2026), China Daily AI Market Analysis, Finextra Luckin Coffee Implementation, SCMP Tencent WeChat Integration, ByteDance Doubao Coverage (36Kr), Baidu ERNIE Bot (Baidu PR), Google UCP Developer Guide, OpenAI ChatGPT Shopping

  • Agentic Commerce
  • AI
  • Ecommerce
Eugene Levitin
Eugene Levitin

CEO, Ivinco

Building Ivinco since 2009 — a Kubernetes consulting firm with 20+ senior engineers managing 1,350+ servers worldwide. Currently exploring how AI agents are reshaping e-commerce infrastructure.