Wuzhen supplied the technology entrance; the CIIE supplied the market entrance. China is drawing a "tradable, trustworthy, and inclusive" roadmap for AI.
1. A turning point: demand is exploding, but the bottleneck has moved
An easy-to-miss signal is that AI's usage intensity has crossed the threshold of "demonstration."
According to the China Internet Development Report 2026, China's average daily token invocation reached 140 trillion in March 2026, growing more than a thousandfold in two years; by June 2026, 924 generative-AI services had been filed and 573 applications or functions registered. Together, these figures say one thing: AI has moved from "what it can do" demonstrations to infrastructure used every day.
Yet even as demand explodes, the bottleneck has relocated. At the CIIE, Dun & Bradstreet made a widely quoted remark: "a trusted data layer is the real AI infrastructure." The weight of this sentence lies in its implication: once model capability ceases to be scarce, what decides whether AI lands becomes whether the data is trustworthy, traceable, and governable.
This is the first judgment: AI's competitive focus is shifting from "model capability" to "landing capability."
2. Three hurdles of landing: tradable, trustworthy, inclusive
If "landing" is the main battleground of AI's second half, the problem can be decomposed into three questions.
First, can it be traded? AI's value ultimately has to land on being "bought, installed on a production line, and regulated." At the CIIE, Schneider Electric exhibited a "native power-management agent" that proactively detects anomalies and proposes dispositions; Siemens brought an end-to-end humanoid-robot solution. These are not technology demonstrations but commercial offerings that can be ordered and deployed. As AI shifts from "exhibit label" to "industrial foundation," its evaluation yardstick likewise shifts from "how many parameters" to "can it be traded."
Second, can it be trusted? The foundation of trust is data. The China report's figures—over 7,000 advanced smart factories, and the digital economy's core industries accounting for 10.5% of GDP—essentially say that AI has entered real production processes, where the requirement for data trustworthiness is far higher than on the consumer side. If the data cannot be trusted, even the strongest model will not be used on a production line.
Third, can it be inclusive? The World report offers two seemingly contradictory judgments at once: AI's contribution to the global economy is projected to reach USD 22.3 trillion by 2030 (3.7% of global GDP), even as AI is giving rise to a new "intelligence divide" that widens the digital gap between countries. The pie is growing, but the gap in how it is divided is growing too. Inclusiveness is an unavoidable question in any landing narrative.
These three questions—tradable, trustworthy, inclusive—are not three separate matters but three facets of the single act of "landing."
3. Two entrances: Wuzhen's technology, the CIIE's market
Notably, the two events each answered half of the problem.
Wuzhen supplied the "technology entrance." The summit's biggest signal this year was "open source" appearing in the theme for the first time. This marks a shift in AI-governance narrative from "regulate" (whose responsibility, who supervises) to "share" (distributing models, capabilities, and tools). Alongside it, "physical AI / embodied intelligence" became the main thread—54 countries' 671 enterprises brought more than 1,000 new technologies, and over a hundred robots entered industrial, eldercare, and entertainment scenarios; the proposals of "AI for Science" and an "agent internet" further elevate AI from a tool to a "research collaborator." Wuzhen answered the question of "how technology is shared and its threshold lowered."
The CIIE supplied the "market entrance." The ninth CIIE alone achieved USD 85.45 billion in intended transaction value on an annualized basis, up 2.3% year on year and a record high; across the first eight sessions, cumulative intended transaction value exceeded USD 580 billion—the most tangible measure of the "market entrance." From May 1, 2026, China has applied zero tariffs to 53 African countries with diplomatic relations; the CIIE upgraded its Asia-Africa product zone, expanded free booths for least-developed countries by 17%, and Côte d'Ivoire attended for the first time as an official delegation. Meanwhile, the "debut economy" and the "exhibitors turned investors" mechanism upgraded "selling goods" into "combined entry via investment, capacity, and standards." The CIIE answered the question of "how markets get in and their threshold lowered."
The two entrances together form a complete proposition: open-source technology plus open markets are the twin wheels of AI landing.
4. Core judgment: the competitive yardstick shifts from "technical capability" to "tradability"
Combining the three questions and the two entrances, my core judgment is this:
AI's competitive yardstick is shifting from "who has the stronger technology" to "who can turn AI into a tradable, trustworthy, and inclusive commodity."
Here, "tradability"—whether an AI capability can be purchased, deployed into real scenarios, and held subject to regulation and accountability—replaces "parameter scale" as the new yardstick. The rationale is threefold. First, model capability is diffusing rapidly through open source, so a pure capability lead is hard to sustain as a durable advantage. Second, tradability, trust, and inclusiveness all depend on market depth and institutional conditions and cannot be achieved through single-point breakthroughs. Third, governance rules are now running in parallel with strategic competition; whoever first closes the loop of "landable + governable" holds the initiative for the next phase.
If this judgment holds, its corollary is clear: for firms, what matters is not "having a large model" but "having trustworthy data, open markets to enter, and scenarios that can be purchased."
5. Practical implications
For firms going global: zero tariffs, the debut economy, and "exhibitors turned investors" together point to a trend—the market's center of gravity is tilting toward the Global South, and the mode of entry is shifting from pure export to a combination of investment, capacity, and standards.
For AI service and consulting practitioners: three entry points are most worth pursuing—(1) the trusted data layer (the real bottleneck of AI landing is not the model); (2) scenario-based landing of physical AI / embodied intelligence (replicable cases have emerged in industry, eldercare, and inspection); (3) the "boundary management" of open source (balancing sharing-collaboration against risk-control is a natural topic for enterprise AI governance).
For organizational managers: AI's second half is, at bottom, an organizational question—how to make data trustworthy, transactions achievable, and responsibility settleable. This is of a piece with the core question of human-AI fit: not "whether humans will be replaced," but "how humans and AI each bear responsibility and land together."
6. Conclusion
AI's first half was a "capability narrative"—whose model is bigger, whose compute is stronger. The second half is a "landing narrative"—who can turn AI into a tradable, trustworthy, and inclusive commodity.
What Wuzhen and the CIIE offered are two halves of the same answer: technology must be open-sourced, and markets must be opened. Connecting the two halves is the word most frequently invoked yet most easily underestimated—trust.
Models will keep growing larger, but what is genuinely scarce is the ability to turn AI into a tradable, trustworthy, and inclusive commodity.
References
- Cyberspace Academy of China. (2026). China Internet Development Report 2026. Released at the WIC Wuzhen Summit, 2026-11-04.
- Cyberspace Academy of China. (2026). World Internet Development Report 2026. Released at the WIC Wuzhen Summit, 2026-11-04.
- CCTV News. (2026-11-04). Coverage of the release of the 2026 WIC blue papers.
- Science and Technology Daily. (2026-11-04). Coverage of the blue papers' core data.
- Cyberspace Administration of China. (2026-11-04). Release notice of the WIC blue papers.
- China International Import Expo. (2026-11). Official information on the 9th CIIE (zero tariffs, debut economy, etc.).
- World Internet Conference / Guangming Online. (2026-11). Data on the Wuzhen Summit "Light of Internet" Expo.