What happened
China's Ministry of Industry and Information Technology has issued an AI small and medium enterprise entrepreneurship support plan covering 2026 to 2028, according to a report from Securities Daily carried by East Money.
The plan aims to cultivate a large group of AI startups with strong innovation vitality and growth potential in key areas such as industry applications, data services and intelligent computing power.
Over the three-year period, the plan targets more than 10,000 newly cultivated technology and innovative SMEs, over 2,000 specialized and sophisticated 'little giant' enterprises, and the emergence of a number of gazelle and unicorn companies.
The support plan addresses inclusive computing power, incubation platforms, the open-source ecosystem and inclusiveness toward entrepreneurial entities, building what the report describes as fertile ground for AI SMEs.
Why it matters
The policy signals that high-quality development of the AI industry depends not only on leading firms tackling foundational large models, but also on numerous smaller companies deepening industry applications, the report argues.
Small and medium enterprises are described as close to the market and familiar with detailed industry processes and operational pain points, making them essential to the 'last mile' of AI deployment.
Rather than pursuing general-purpose large models, these firms are portrayed as rooted in vertical scenarios such as manufacturing, metallurgy, healthcare and industrial inspection, developing smaller models and industry solutions.
The report frames the AI ecosystem as a layered, collaborative system in which leading companies build the computing and model foundation while SMEs act as scenario converters, industry adapters and local service providers.
Key facts
The support plan covers the period 2026 to 2028.
It targets more than 10,000 newly cultivated technology and innovative SMEs.
It targets more than 2,000 specialized and sophisticated 'little giant' enterprises.
Key areas named include industry applications, data services and intelligent computing power.
The plan spans inclusive computing power, incubation platforms, the open-source ecosystem and inclusiveness toward entrepreneurial entities.
What to watch next
The report identifies high computing costs, difficulty obtaining data and talent shortages as three common barriers for AI SMEs, and points to public computing platforms, computing voucher subsidies and inclusive model-calling services as ways to lower them.
It also calls for stronger supply-demand matching, suggesting leading industrial chain firms could use industrial parks and industry associations to build matchmaking platforms and open application scenarios for pilot opportunities.
A third area to watch is capital: the report argues investment logic is shifting from chasing concepts toward real deployment revenue and industry depth, and calls for patient capital covering seed, startup and growth stages.