On 26 August 2025, the State Council issued the Opinions on Deepening the Implementation of the “AI Plus” Action. The document consolidates previously scattered AI policies into a single implementation line and sets quantitative targets.
Three phase goals
The Opinions set three milestones:
By 2027, AI should be widely and deeply integrated with the six priority sectors, adoption of next-generation smart terminals and AI agents should exceed 70%, the core industries of the intelligent economy should grow rapidly, and AI’s role in public governance should be markedly strengthened.
By 2030, AI should fully empower high-quality development, adoption of next-generation smart terminals and AI agents should exceed 90%, and the intelligent economy should become a major growth pole.
By 2035, China should fully enter a new stage of intelligent economy and intelligent society.
Using “adoption rate” rather than “industry scale” as the core metric is a telling choice. It places the evaluation standard on whether the technology is actually used, continuing the “large-scale commercial application” theme raised at the State Council executive meeting in July. See the State Council executive meeting on the AI Plus opinion.
Six priority sectors
The Opinions set out tasks across six directions: science and technology, industry, consumption, livelihoods, governance and global cooperation. This division covers the main scenarios from production to daily life and from domestic to international.
Notably, “livelihoods” and “governance” are listed separately. That means AI applications in healthcare, education and government affairs will receive policy attention equal to industrial applications. See the AI Plus healthcare implementation opinion and the guideline on deploying government AI large models.
‘AI agents’ enter policy vocabulary for the first time
The Opinions place “next-generation smart terminals” and “AI agents” side by side as the objects of the adoption-rate target. This is the first time “AI agents” appear in a State Council-level quantitative goal.
Agents mean AI moving from generating content to executing tasks, and their compliance implications are entirely different: an agent may autonomously call tools, access data and initiate transactions, so responsibility and permission boundaries must be redefined. See the national standards on agent interconnection.
Computing power as the backbone of the safeguards
The Opinions call for strengthened coordination of intelligent computing power, support for breakthroughs in AI chips and the cultivation of an enabling software ecosystem, faster technical breakthroughs and engineering deployment of ultra-large-scale intelligent computing clusters, optimisation of the national layout of intelligent computing resources, and improvement of the national integrated computing network.
This line connects with existing computing power policy. See the national integrated computing system.
Where businesses can start
First, map your business onto the six priority sectors and judge which directions will receive both policy and demand support.
Second, plan ahead for agent-related capabilities and compliance frameworks, including permission management, operation logs and responsibility allocation.
Third, track sector-level implementing rules. The Opinions are usually delivered through supporting documents from ministries and localities. See the AI Plus manufacturing action plan.
The significance of the Opinions is that they turn “AI Plus” from a slogan into an implementation plan with goals, sectors and safeguards. For businesses it is both a demand signal and a compliance agenda.
This article is based on the State Council’s Opinions on Deepening the Implementation of the ‘AI Plus’ Action and public information and is for reference only; it does not constitute legal advice.