This morning, the 2026 World Intelligent Industry Expo opened at the Tianjin National Convention and Exhibition Center. The exhibition area covers 130,000 square meters, with over 700 enterprises and 6 themed pavilions.
A highlight I hadn't anticipated: embodied intelligence has its own dedicated pavilion for the first time this year. It's not squeezed in with robots, nor is it a corner of the AI pavilion—it's an entire hall. JD.com has moved its "Robot Town" here, and Unitree's robotic dog is making coffee for visitors—picking up the cup, placing it on the conveyor belt, turning around to continue with the next one. The movements aren't particularly fast, but they are steady.
If this had happened last year, people might still have seen it as a novelty. But now, few doubt one thing: AI is growing a body. And the pace of this is faster than everyone predicted in 2024.
Policies were intensively introduced in April and May, and intelligent agents are no longer just a "trend"
In May, the National Cyberspace Administration, the National Development and Reform Commission, and the Ministry of Industry and Information Technology jointly issued the "Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents." This is not a "encourage exploration" type of guiding opinion, but rather provides a formal official definition for AI agents:
"An intelligent agent is an intelligent system with autonomous perception, memory, decision-making, interaction, and execution capabilities."
Note the last two characters: execution. It is not chatting, not Q&A, not text generation—it is execution. Nineteen typical application scenarios are explicitly listed, covering five directions: scientific research, industrial development, boosting consumption, people's livelihood and well-being, and social governance. Scientific research is placed first, and this order itself is a signal.
In the same month, the "AI+" initiative was upgraded from "deepening and expanding" at the Politburo meeting in late April to "full implementation." The difference of two characters implies that the pilot phase has ended and large-scale deployment has begun.
Putting these together: a national-level expo opens today, with embodied intelligence as a standalone pavilion; a three-department document provides official definitions and a scenario list; "AI+" shifts from pilot to full implementation. The convergence of these three threads in the same month is no coincidence.
OpenClaw ignited "Operation AI," and all the major players have entered the fray.
The trigger for this change was an open-source AI agent framework called OpenClaw (Lobster), which can see the screen, click the mouse, and work automatically. This in itself is not surprising—what is surprising is the speed at which it exploded.
Baidu's Robin Li proposed a concept called DAA (Daily Active Agents), which is meant to replace DAU from the mobile internet era. His original words were: "The rise of agents shows that what users truly pay for is not 'whether you can do it,' but 'whether you can help me get things done.' This marks AI's transformation from a chat tool to a digital employee."
On Alibaba's side, Tongyi Qianwen launched a "citation" feature in March this year — a bright green "credible" label for reliable sources, and a red warning for vague sources. On May 11, Qianwen was fully integrated with Taobao, with "ordering takeout in one sentence" backed by the integration of hundreds of small projects. Taobao covers over 300 prefecture-level cities and more than 3,000 districts and counties. At this scale, not leveraging AI would be a waste; leveraging AI is inevitable.
Moon's Dark Side (Kimi) has already produced data in financial and educational scenarios: what used to take securities researchers 2-3 days now takes 2-3 hours; what used to take doctoral students 2-3 weeks for literature review now yields a first draft in 1-2 days. The MoE architecture with trillion-level parameters activates 32 billion parameters per inference, with 384 domain experts—not everyone may understand the technical parameters, but the figure of "2-3 days reduced to 2-3 hours" is enough to make any enterprise's CIO restless.
360 AI Security Research Institute released a report titled "When AI Has 'Hands and Feet,' Enterprise Security Boundaries Must Be Rebuilt." The core message is: In the past, AI security issues mainly involved preventing "nonsense," but now it is about preventing "misbehavior." Two approaches: using AI to enhance traditional security protection, and allowing uncertain tasks to be executed under security constraints.
Tencent, ByteDance, and Zhipu have also entered this track, but their public moves are not as clear as those of the companies mentioned above. However, this is not important—what matters is that when everyone is running in the same direction at the same time, that direction is probably right.
The capital market is also repricing
According to Securities Times Data Treasure statistics, the average daily trading volume of AI agent concept stocks since May has been 7.84 billion yuan, a month-on-month increase of 30%. The average daily trading volume of 22 stocks doubled month-on-month. The average daily trading volume of Nanwei Software surged 688% to 505 million yuan — this company operates the full chain of "computing power + large model + agent" and has been certified by the China Academy of Information and Communications Technology.
17 concept stocks are predicted by more than 5 institutions to have net profit growth rates exceeding 20% this year and next. Interestingly, 10 have retraced over 20%—Hehe Information has retraced 50% from its January high, and Dian Diagnostics has retraced 47%.
Hehe Information's AI product is called Docflow, which focuses on automated document processing — collection, parsing, classification, extraction, and review, positioning itself as the "document infrastructure for enterprise digital transformation." This kind of thing may not sound as sexy as general-purpose large models, but it can generate revenue.
Dean Diagnostics is more direct: tertiary hospitals and county-level medical alliances are already paying to use their AI pathology and intelligent diagnostic products.
The market's reaction this time illustrates a principle: capital is no longer paying for "AI hype"; it is looking for companies with paying customers, specific scenarios, and delivery capabilities.
852 billion: China AI agent market size forecast for 2028 (554 billion in 2023, CAGR 72.7%)
What it means for enterprise CIOs
Connecting a few things from today, the thread I can see is this:
First, the infrastructure is already in place. The three-department document provides a compliance framework, Robin Li's "DAA" offers a measurement yardstick, and the products at the Smart China Expo are right before your eyes. 2026 is not a year of "whether to use AI," but a year of "how to use AI without being left behind."
Second, embodied intelligence is not a gimmick. The fact that it has its own dedicated hall indicates that the industry chain is maturing—from chips to sensors to motion control to the AI brain, the supply chain can now support an independent industry exhibition. Manufacturing companies should start paying attention; it's not about buying robots, but about closing the AI decision-making loop on the production line.
Third, security is no longer optional. When agents shift from "talking" to "acting," errors are no longer just embarrassing — they mean real financial losses. The "security constraint framework" in the 360 report and the "full-chain security requirements" in the three-department document both point to the same thing: before enterprises adopt AI, they must first adopt governance.
Fourth, taking small steps and running fast is more practical than achieving everything at once. Neither this exhibition nor the policies point to a "comprehensive transformation"—the policies mention gradually advancing 19 scenarios, while Kimi and Tongyi Qianwen both start from a single scenario. Rather than devising a grand AI strategy, it is better to find a specific scenario that can show results within a month and get it running first.
The expo runs until the 31st. In the following days, there will be signings, releases, and bilateral meetings. But the first clear takeaway today is this: the shift of AI from "can it work" to "can it actually get things done" is not something for next year, nor for next quarter—it's happening this month.
Data sources for this article: Xinhua News Agency (reported on May 26, 2026), Securities Times · Data Treasure (May 28, 2026), China News Service (May 28, 2026), and the "Implementation Opinions on the Standardized Application and Innovative Development of Intelligent Agents" (May 2026) by the Cyberspace Administration of China / National Development and Reform Commission / Ministry of Industry and Information Technology.
Disclaimer: This article does not constitute any investment advice. All views expressed in the article are solely those of the author.
