Shanghai Issues 15th Five-Year Plan for Software and Information Services, Targeting RMB 4 Trillion by 2030

On August 11, the Shanghai Municipal Commission of Economy and Informatization issued the 15th Five-Year Plan for the Development of Shanghai's Software and Information Services Industry. The plan proposes that by 2030, Shanghai's software and information services industry will be built into a power source for economic growth, a main front for AI-enabled applications, and a bridgehead for participating in global competition. In terms of overall industry scale, the industry will strive to reach RMB 4 trillion, with industry value added exceeding RMB 1.1 trillion. In terms of industry quality and efficiency, a number of breakthrough results will be achieved in key areas such as artificial intelligence and key software, with companies generating revenue exceeding RMB 10 billion increasing to 35, forming a number of quality enterprises with dominant influence over the industrial ecosystem and emerging enterprises with potential leading influence. In terms of industry development structure, the proportion of high-end software, digital content, digital-intelligence services and other industries will be further increased, and a number of industrial bases and regional industrial clusters with international competitive advantages will be cultivated and formed.

The plan deploys multiple tasks around giving play to the leading role of artificial intelligence. In the digital-intelligence leap of the internet industry, platform enterprises in e-commerce, social networking, local life services and other fields will be encouraged to intelligently restructure their commercial foundations, accelerate the development and application of industry large models and intelligent agents, promote the transformation of traditional internet from traffic operation to value management, and incubate new paradigms such as AI-customized consumption, immersive social spaces and all-domain intelligent butlers. In the digital-intelligence reshaping of the software industry, the AI-oriented development of software products will be promoted, accelerating the intelligent upgrade of high-value scenarios such as industrial drafting, engineering design and production scheduling, and developing AI-native intelligent software to expand the scenario-based service boundaries of software products. In the digital-intelligence upgrade of the digital content industry, various innovation entities will be supported in developing AI-assisted creation tool systems, building intelligent creation platforms that run through the entire chain of script generation, original painting, modeling and rendering, character generation, automatic storyboarding, dubbing and scoring, and promoting the multi-form linked development of quality IP across games, animation, film and television, online literature and physical derivatives.

In the cybersecurity industry, the plan requires promoting the deep integration of artificial intelligence with the entire chain of security defense, detection, response and governance, developing distinctive solutions such as intelligent computing cloud security, AI firewalls, security ranges and intelligent agent identity management, and supporting third-party institutions in expanding service capabilities such as model security testing, content compliance evaluation and AI security planning consulting. In the intelligent computing cloud industry, quality cloud service providers will be guided to build lightweight, high-throughput AI inference foundations, build comprehensive intelligent computing public platforms integrating computing scheduling, model incubation and application adaptation, support cloud service providers in transforming into providers of computing power plus development tools plus model services, and create benchmark application scenarios for vertical models in key areas such as finance, manufacturing, education, healthcare, culture and tourism, and urban governance. In the fintech industry, financial industry large models will be developed, innovative products such as intelligent investment advisory agents and AI compliance risk control systems will be developed, and innovative applications of digital RMB in scenarios such as the metaverse and supply chain finance will be promoted.

In the digital-intelligence transformation of industry, a model-data resonance action will be implemented, building high-quality industrial corpus datasets and creating industry models, and a scenario partner plan will be implemented to incubate plug-and-play vertical intelligent agents and industrial apps.

The plan lays out future-oriented emerging tracks. In the agent economy, core technologies such as general intelligent agents and multi-agent collaboration will be tackled, innovation entities will be supported in developing cross-domain, multi-task, self-planning general intelligent agent software, industry-specific intelligent agents will be cultivated for fields such as industrial manufacturing, professional services, livelihood services and urban governance, and new business models such as agent-as-a-service will be developed. In intelligent terminals, around new-generation intelligent terminals such as embodied intelligence, low-altitude equipment, intelligent wearables and brain-computer interfaces, forward-looking deployment of research and development in dedicated operating systems, main control software and high-end embedded software will be carried out, tackling core technologies such as three-dimensional spatial computing, multimodal cognitive perception, anthropomorphic motion control in complex environments and highly agile human-machine collaborative interaction. In world models, world models with physical cognition, real-time deduction and autonomous iteration capabilities will be tackled, key technologies such as multi-physics coupling simulation, high-fidelity real-time simulation and high-quality data synthesis will be broken through, and the application of technological achievements in key scenarios such as embodied intelligence, aerospace and biomedicine will be promoted.

In AI technology research, the plan proposes tackling next-generation model architectures, promoting exploration of multiple technical routes based on non-Transformer architectures such as state space models, recurrent neural network variants and liquid neural networks, and laying out frontier foundational model technology systems such as physical intelligence, world models, quantum intelligence and brain-inspired intelligence. Ultra-large-scale intelligent computing cluster networking technology will be tackled, improving intelligent computing hardware supply capabilities around core links such as high-performance computing chips, quantum chips, high-speed optical interconnection, high-bandwidth memory and heterogeneous servers, and promoting deep integration of self-developed chips with mainstream large models. In key software development, continued efforts will be made in underlying technology research such as OS kernels, graphics engines, SQL engines and high-concurrency processing, focusing on improving the performance of basic software such as operating systems, databases, middleware and tool software, establishing a software and hardware adaptation verification system, and breaking through core technologies such as geometric modeling, constraint solving, mesh generation and physical simulation, strengthening research and development and innovation in industrial software such as computer-aided design, computer-aided analysis and electronic design automation.

The plan proposes implementing a hundred-thousand-ten-thousand intelligent computing cluster project, forming a tiered supply system of large clusters, small clusters and edge computing power in collaboration. In areas such as Songjiang, Lingang and Qingpu, 100,000-card-scale ultra-large intelligent computing clusters will be laid out and built; in areas such as Baoshan, Pudong and Jiading, 1,000-card-scale intelligent computing clusters will be laid out and built; and traditional data centers and information and communication computer rooms will be guided to transform into 100-card-scale edge intelligent computing centers, meeting the ultra-low-latency computing power needs of enterprises and individuals. In open-source ecosystem building, relying on open-source platforms such as the Baiyulan Open Source Research Institute, OpenLoong and OpenDataLab, open-source projects in intelligent software, corpora, algorithms, models, intelligent agents, and humanoid robot software and hardware will be promoted for first release in Shanghai, and open-source benefit distribution mechanisms such as paid open source and dual licensing will be promoted.