On 29 April 2026 the Ninth Digital China Summit opened in Fuzhou under the theme “Accelerating Digital-Intelligence Technology Innovation, Deepening Digital China Construction”. The summit hosted more than 50 dialogue sessions, a Digital China innovation competition and an on-site experience zone, alongside special activities. The AI track was a core focus.
A stage judgment: from generation to execution
A judgment repeated through the summit is that AI is moving from “talking” to “doing”. The phrasing is colloquial but points at a real technological and governance turn.
In the “talking” stage, AI’s core capability is content generation, and risk concentrates at the output end — whether content is true, compliant and labelled. Entering the “doing” stage, AI gains planning and execution capability, able to call tools, operate software and drive devices. Risk then expands from the output end to the action end: who authorises, who is liable, how to audit.
Two main lines on the industry side
The first is agents. Agents are seen as the carrier connecting the digital world with real scenarios, moving AI from single-point breakthroughs to system coordination and from passive response to autonomous execution. The corresponding governance requirements are more concrete — see our analysis of the three ministries’ opinion on AI agents.
The second is embodied intelligence and terminals. Robots and smart terminals were shown in concentration, as AI moves from software to hardware. Terminal-side compliance questions follow: how device-collected data is handled, how on-device models are filed, and what notice duties apply to consumer products — see our analysis of the “AI plus consumption” implementation opinion.
The link between Digital China and data elements
One main line of the summit is digital-intelligence technology innovation; another is the overall advance of Digital China. The connecting point is data. AI’s capability ceiling increasingly depends on accessible high-quality data, while the release of data value depends on the improvement of property-rights, circulation and security systems.
This is consistent with the National Data Administration’s parallel push on data property-rights registration and high-quality industry datasets — see our analysis of the draft guide on data property-rights registration and the action plan for high-quality industrial datasets.
What it means for AI companies
First, as product form changes, the compliance framework must follow. Moving from content generation to task execution requires permission management, behaviour auditing and responsibility-allocation mechanisms — see our AI law practice.
Second, scenario deployment requires industry knowledge. As AI integrates with manufacturing, energy and healthcare, compliance questions take on sector attributes and must be handled under sector rules — see our multi-jurisdiction service.
Third, data assets need managing. Companies participating in the data-element market should build data ledgers and source evidence — see our data-compliance service.
The summit’s value is giving direction, not answers. For AI companies, seeing the “talking to doing” judgment clearly means thinking ahead about one question: when the product starts doing things for users, who bears the responsibility.
This article is based on public information concerning the Ninth Digital China Summit opened on 29 April 2026. It is for general reference only and does not constitute legal advice.