Guangdong's AI policy over the past few months has formed a chain: in April, the region-wide application action plan was issued; afterward came the "Yue Zhi Fu" empowerment platform and resource selection; and now the funding end has been added—the 2027 provincial-level artificial intelligence technology breakthrough project has been included in the project pool, Document No. Yue Gongxin Shuzi Chanye Han [2026] No. 68.
Simply put, this is the province preemptively "reserving a spot" for next year's special AI funding: projects first enter the database, and only after being included are they eligible to be listed within the scope of provincial special fund support.
What is supported
Two directions.
First, tackling common AI technologies, focusing on technology and platform categories.
Second, tackling AI industry scenario applications; the notice explicitly named the industries: software and information, electronic information, machinery, textiles, shipbuilding, home appliances, metallurgy, medical devices, ceramics, toys, and luggage.
Note the composition of this industry list—half of it consists of Guangdong's traditional advantaged manufacturing industries. The appearance of industries such as toys, ceramics, and bags and luggage in a provincial-level AI special document shows that what the provincial government wants is not just a few high-end, sophisticated demonstration cases, but real cases of AI being implemented across all industries and sectors.
Three key points in the rules
First, the city recommendation system. Enterprises cannot directly submit materials to the province; they must report to the industry and information technology bureau of the city where they are located, and the city shall review them and issue a recommendation opinion. The notice is very blunt: in principle, projects that have not been reviewed by the city shall not be included in the scope of provincial special fund support.
Second, time is very tight. The forwarded notices from Qingyuan, Guangzhou, and other places all require that the application form, budget form, and recommendation summary table be submitted to the municipal Industry and Information Technology Bureau by November 11, with seven paper copies plus an electronic version. Reporting upward also requires time for district/county preliminary review and municipal bureau review, so the actual window for action is just these few days.
Third, being included in the pool does not mean getting money, but not being included definitely means no money. Inclusion in the pool is a qualifying round; next year's special funds will only select projects from the pool.
Which project matches?
Based on the document criteria and practical implementation experience, the types of projects that are relatively promising in manufacturing enterprises are roughly the following categories:
AI visual inspection in quality inspection (including industries that traditionally rely on manual visual inspection); intelligent scheduling optimization in production scheduling; predictive maintenance for equipment operation and maintenance; demand forecasting in supply chain.
What must be clearly stated in the proposal is the key-problem-tackling nature—how it differs from buying off-the-shelf generic software, where industry know-how is accumulated, and the quantifiable indicators for expected cost reduction and efficiency gains. There is an unavoidable precondition here: the precondition for AI application breakthroughs is data availability. Enterprises whose production line data is still on paper documents and whose quality records have not yet been structured should not rush to apply for AI; even if they do, they will not be able to write a persuasive proposal.
A reminder for each of the two types of enterprises
Manufacturing enterprises that have mature AI application scenarios and whose data foundation is in place: immediately contact your local industry and information technology bureau to confirm the recommendation process and material requirements. Once the 11-11 deadline passes, there will be no further opportunity.
Enterprises with weak data foundations but an interest in AI: save the industry scenario list in Document No. 68 as a reference for 2027 project planning. This year, build a solid foundation in ERP and business data; when such windows come again next year, you will be prepared.
One-sentence summary: This is the admission ticket for the funding side; the gate closes on November 11.
