Shanghai's economy and IT commission published a new smart manufacturing action plan on October 27, making AI integration across the whole manufacturing chain its main line: by 2028, all key manufacturers in the city should run smart factories, 8,000 humanoid robots should be working on production lines, and AI-empowered smart manufacturing equipment should exceed RMB 200 billion in output.

__FIG__

1. A "5-4-3-2-1" roadmap

The plan follows the principle of "scenario-driven, model-powered, equipment-upgraded, capability-spillover". Its quantified goals form a 5-4-3-2-1 framework: 500 advanced, 50 excellent and 5 leading smart factories; a 4-tier cultivation system; 3 tiers of integrated service providers; 2 platforms for supply-demand matching and overseas expansion; and 1 pilot-testing platform for smart manufacturing.

Key targets by 2028

- 1,600 basic-level and 500 advanced-level smart factories;

- Robot density of 600 units per 10,000 workers, 8,000 humanoid robots in factories;

- 10 vertical LLMs, 300+ high-level agents, 2,000+ AI application scenarios;

- RMB 200B equipment output and RMB 100B system-integration output.

2. 2,000+ AI scenarios on the shop floor

Basic-level factories must reach 100% AI adoption, while advanced, excellent and leading factories should apply AI in 10%, 30% and 70% of scenarios respectively. Ten vertical LLMs will be built for high-end equipment, shipbuilding and industrial software, with 300+ agents covering design assistance, adaptive production and predictive maintenance. A "models policing models" security framework adds 10 benchmark scenarios such as AI security operation centers.

3. From pilot lines to chain-wide upgrades

Shanghai will build 10 model enterprises with dense agent deployment, support chain-leader companies in upgrading their suppliers, and promote lightweight, low-cost solutions for SMEs. For solution providers, vertical LLMs and industrial agents are now a policy-backed market. NineZenith's Zenith-Act industry LLM supports private deployment, helping manufacturers build agents for scheduling and predictive maintenance that fit this build-out cycle.

(Compiled from public reports by 21st Century Business Herald and the Shanghai Municipal Commission of Economy and Informatization)