Chinese Premier Li Qiang has again linked artificial intelligence to manufacturing, but his latest Shanghai visit focused on a more practical part of the policy agenda: the shared services that can help companies use the technology. During an inspection of an AI application pilot test base on September 22, Li called for more manufacturing platforms covering industrial design, pilot testing, and intelligent operation and maintenance.
The official gov.cn and Xinhua account places the visit on September 22, while SCMP reported it the following day. The two accounts converge on the central point. Li urged a deeper combination of China’s manufacturing strengths with digital and intelligent technologies and called for further implementation of AI plus manufacturing. SCMP says the proposed platforms are intended to help companies, particularly small and medium-sized enterprises, upgrade operations more efficiently and at lower cost.
The remarks echo a broader call Li made in May for faster AI adoption in advanced manufacturing, covered by EastFrontier in China’s earlier push for broader industrial AI use. The September visit is not a new law, subsidy, procurement program, or company launch. Its significance is more specific: it identifies a set of services that could make AI adoption easier for factories that cannot build every capability themselves.
The manufacturing services Li highlighted
The official report says Li visited an AI application pilot test base and inspected special-purpose and industrial robots, including uses in complex scenarios. SCMP identifies the functions he emphasized as industrial design, pilot testing, and intelligent maintenance. Gov.cn uses closely related language, adding intelligent operation and maintenance and the need to address common manufacturing challenges.
Those are not abstract categories. Industrial design can involve creating and testing a new component or product configuration. Pilot testing is the stage where a process is tried on a limited scale before it is deployed across a factory. Intelligent maintenance can mean using sensor data and software to identify a likely fault before it creates a long shutdown. AI can contribute to each area, but most manufacturers need data, integration, and technical support before they can use it reliably.
That is why a shared platform model matters. A large company may have the budget to build its own machine-learning team, industrial data system, and testing environment. A smaller supplier often does not. If service platforms can provide tools, test facilities, reference data, or technical support, they could lower the cost of experimenting with AI in a setting where a failed trial can disrupt production.
The official report does not name the operator of the Shanghai base, the companies involved, a particular model, or a funding amount. It should therefore not be read as an announcement of a new national facility. The policy signal is instead about the type of capability Li wants more manufacturers to access.
From broad policy to shared services
China has spent years encouraging industrial digitalization and intelligent manufacturing. The country’s AI policy language has often stressed integration across factories, supply chains, and equipment. The risk with broad adoption goals is that they can sound complete before companies know how to make them work in a specific production line.
Li’s focus on design, pilot testing, and maintenance turns that broad goal into a service-layer question. A factory may not need a general-purpose chatbot. It may need help identifying a defect pattern, changing a production schedule, testing a new material, or deciding which maintenance event deserves a technician’s attention. Those tasks require tools that connect to physical processes and have a measurable operational use.
This is consistent with the earlier data on China’s AI-driven manufacturing expansion, but the latest message is more restrained than a claim of completed transformation. Li did not announce a nationwide result. He called for platforms that can help companies tackle common problems and critical links in their operations.
That distinction is important for small and medium-sized companies. Many have useful production data but lack standardized systems, specialist staff, or the time to build a custom application. A shared platform could potentially make a pilot less expensive and reduce the chance that a project stalls after an initial demonstration. SCMP’s language about more efficient and lower-cost upgrading describes the policy goal, not a measured savings figure.
Robots, future industries, and the capital behind them
The official account also shows that the visit extended beyond one factory application. Li inspected special-purpose robots and industrial robots at the AI pilot test base. At another science-and-technology innovation and incubation platform focused on future industries, he viewed products related to brain-computer interfaces and quantum computing.
Those stops reveal a wider policy frame. AI manufacturing is not being treated as a standalone software issue. It sits alongside robotics, new computing approaches, and the infrastructure needed to turn technical research into products. The official report says Li noted that future industries often have long growth cycles and high market risks, and that he called for more venture capital and patient capital.
That is not a financial commitment to a particular firm. It is an acknowledgement that industrial technologies can take longer to validate than consumer applications. A factory tool may need to work through safety tests, integration with existing machines, and several production cycles before a buyer will trust it. Investors and policy platforms that expect an immediate consumer-style growth curve may struggle with that reality.
The service-platform idea could help bridge some of that gap. A reliable test environment may let a company validate an AI tool before risking a live line. A maintenance platform may make the value of a system easier to measure. A design tool may reduce iteration time. None of those outcomes is guaranteed, but they are more concrete than a general call for factories to use more AI.
The Shanghai visit therefore offers a useful update to China’s industrial-AI agenda. Li Qiang’s September 22 remarks centered on the mechanisms of adoption: design, testing, and maintenance services that can support manufacturers, especially smaller ones. The next evidence to watch is not another broad policy slogan. It is whether those platforms become identifiable, accessible tools that factories can use and evaluate in their own operations.