On 21 August 2026, China’s cyberspace portal published the Action Plan for Promoting the High-Quality Development of Cyberspace Enterprises (2026-2030). The document positions cyberspace enterprises as an important force in building a strong cyber nation, and around the theme of “cultivation and service” arranges five blocks — enterprise cultivation services, innovation capability, industrial upgrading, overseas market expansion, and safe and orderly development — with a number of specific actions.
For AI companies, the value of this plan lies in the fact that it gives two things at once: policy support to be pursued, and compliance baselines to be held.
Tiered cultivation: enterprises in three categories
The first task is to build a database of quality cyberspace enterprises and cultivate them by tiered and classified standards, in three categories:
Ecosystem-leading — enterprises with the capacity to lead on technology and standards, integrate factors and markets, and build industries and ecosystems; industry-leading — leaders in a niche field able to lead technology-standard setting and supply-chain integration; high-growth and innovative — SMEs with strong technological attributes focused on differentiated innovation in a niche, aiming to cultivate more gazelle and unicorn enterprises.
The document expressly calls for integrating resources such as investment and financing institutions, accounting firms, law firms and IP service providers to build an incubation and cultivation service mechanism. This is a clear participation signal for professional service providers, and means enterprises selected into the cultivation database will find it easier to connect with compliance and IP resources.
Innovation capability: AI key technology singled out
In the innovation-capability action, artificial intelligence is singled out. The document requires support for enterprises to strengthen research on AI key technologies, develop high-end AI chips, break through key technologies for high-performance training clusters, improve the performance of foundation models such as large language, multimodal and world models, and lay out frontier technologies such as AI agents and embodied intelligence.
Supporting this are IP tools: comprehensively using IP priority examination and expedited examination to strengthen service support, strengthening the construction of industrial IP operation centres in the cyberspace field, and promoting patent conversion of scientific and technological achievements. For AI companies whose core assets are models and algorithms, the value of patent and open-source compliance planning is further highlighted; see our IP Protection service page.
Industrial upgrading: manufacturing and digital consumption are the main arenas
The industrial-upgrading action gives specific directions. In manufacturing, it encourages enterprises to deploy industry large models and AI agents for manufacturing, build high-quality industrial datasets, and strengthen R&D and application of humanoid robots; in digital consumption, it encourages increasing the supply of AI terminals and wearable devices, and developing new formats such as “AI+” medical services. The document also stresses accommodating the needs of the elderly and people with disabilities, improving the convenience and friendliness of digital services.
These directions map directly to product forms and data-compliance arrangements: industry large models involve industrial data, and terminal products involve on-device personal information processing. For related boundary questions, see our Data Compliance service page.
Going global: from encouraging overseas expansion to building a compliance system
The overseas market expansion action is the most practical part of the plan for AI companies. The document proposes supporting advantageous enterprises in computing power, cloud services and artificial intelligence to “go global” in an orderly way while preventing risks, and expressly calls for “building a service system for enterprises going global”:
strengthening overseas business compliance guidance for cyberspace enterprises, exploring the establishment of overseas business compliance guidance centres for key markets; establishing and improving risk monitoring and early-warning mechanisms, guiding enterprises to respond effectively to unreasonable trade-restrictive measures; building a case database of legal disputes for cyberspace enterprises going global, and encouraging trade and investment promotion agencies, industry associations and chambers of commerce to provide legal services and training.
Overseas compliance is thus brought from “the company’s own business” into the supply of public services — a substantive benefit for AI companies preparing to enter markets such as the EU and Southeast Asia. For related discussion on multi-jurisdiction adaptation, see our Multi-Jurisdiction Compliance and EU AI Act service pages.
Safe and orderly development: compliance baselines are equally clear
The other side of the plan is regulatory requirements. The safe and orderly development action proposes: continuing the special governance of mobile internet applications illegally collecting and using personal information; and continuing comprehensive governance of internet information service algorithms, urging optimisation of algorithm operating mechanisms and management rules, improving review of algorithm-recommended content, strengthening governance of AI technology abuse, and safeguarding the lawful rights and interests of internet users and workers.
On competitive order, the document requires strengthening antitrust enforcement in the cyberspace field, strictly prohibiting disguised “either-or” choices and other conduct that excludes or restricts competition, regulating platform fees, and focusing on regulating excessive subsidies and “involution-style” competition. For AI application companies that capture markets through subsidies, this is a red line to assess in advance.
The document also proposes actively conducting forward-looking legislative research in the cyberspace field and accelerating legislation in areas such as artificial intelligence, countering online violence and the digital economy. This means the supply of AI rules will continue to densify, and companies should plan compliance investment as a continuous process rather than a one-off project.
Three things for AI companies
First, view policy opportunities and compliance obligations on one sheet. The plan provides support in cultivation, funding, IP and overseas services, and also sets requirements on algorithm governance, antitrust and personal information protection; the two need to enter annual business planning together.
Second, arrange overseas expansion early rather than late. Mechanisms such as overseas business compliance guidance centres and the dispute case database are being built; companies can first map out the compliance checklist for target markets and connect directly when services land, reducing the cost of improvised responses. For support on overseas compliance pathways, overseas structuring and cross-border data arrangements, see our Multi-Jurisdiction Compliance service.
Third, watch the application windows for the cultivation database and resource pool. Tiered and classified cultivation means companies must first have a clear technology and compliance profile to be accurately categorised and enter the cultivation service mechanism.
This article is based on the Action Plan for Promoting the High-Quality Development of Cyberspace Enterprises (2026-2030) published on 21 August 2026. It is for reference only and does not constitute legal advice. Specific application should be based on the official document.