Sichuan has introduced a new policy framework, issuing 19 measures to advance the construction of its computing power network. Released on August 14th, the "Several Policy Measures for Promoting the Construction of the Computing Power Network in Sichuan Province (Trial)" outlines a strategic plan to develop computing resources in a scientific and orderly manner, deeply integrating into the national unified computing power network.
Sichuan plays a pivotal role in the national "East Data, West Computing" project and is a key node within the Chengdu-Chongqing national computing power network hub. These 19 measures, developed by the provincial development and reform commission in collaboration with the economic and information technology department, the science and technology department, the energy bureau, and the communications administration, define key tasks and responsibilities. The strategy focuses on building computing facilities with clear demand and appropriate scale ahead of schedule, thereby strengthening the computing power foundation for Sichuan's high-quality digital economy development.
Establishing a "One Zone, One Belt, Two Nodes" layout, the measures specify the role of each region. In the early period of the "15th Five-Year Plan," the central government has prioritized the construction of six major infrastructure networks: water networks, new power grids, computing power networks, next-generation communication networks, urban underground pipelines, and logistics networks. A computing power network connects data centers, supercomputing centers, and intelligent computing centers across different locations via high-speed networks, enabling on-demand usage and making computing power as accessible as electricity. Sichuan recently set a goal to become the western region's computing power hub by the end of 2028, aiming to triple its intelligent computing capacity from the end of the "14th Five-Year Plan" period to 240 EFLOPS, representing over 90% of its total computing power. The new measures define regional roles within this plan: strengthening the Chengdu Plain core computing area, developing the Panxi-Northwestern Sichuan computing-electricity integrated development belt, and cultivating the Southern and Northeastern Sichuan regional computing nodes. The Chengdu Plain core area, encompassing Chengdu, Ya'an, Mianyang, and Deyang, will focus on intensive and large-scale development of computing resources, creating a comprehensive computing hub serving the province, the nation, and countries along the "Belt and Road." The Panxi and Northwestern Sichuan regions, with their abundant clean energy resources, are designated for supporting the layout of integrated "source-grid-load-storage" large-scale computing facilities in Liangshan, Ganzi, Aba, and Panzhihua, forming a computing-electricity belt for non-real-time, transferable computing needs like model training, big data analysis, film rendering, and data backup. In Southern Sichuan, cities like Neijiang and Yibin are supported in building computing facilities for industries such as smart logistics, intelligent brewing, and smart manufacturing, creating a regional node serving Southern Sichuan and surrounding areas. In Northeastern Sichuan, Dazhou is supported in developing computing facilities for energy, chemicals, machinery manufacturing, and modern agriculture, forming a node for that region and its neighboring areas.
Strengthening computing parks to enhance supply and promote universal access, the plan focuses on making computing power a key driver. Computing parks, as core production units of the network, are a priority. Sichuan will optimize the Tianfu data center cluster, build zero-carbon computing parks, and upgrade existing ones. The policy supports Chengdu in establishing advanced computing parks in Chengdu Science City, Shuangliu Economic Development Zone, and Chengdu Chip Valley, an intelligent computing park in Future Science City and the Gaoxin-Pidu Electronic Information Industrial Park, and a smart cloud computing park in the Eastern New Area, including Konggang, Jianzhou, and Jianyang. By integrating hydro, wind, solar, and storage resources from the Panxi-Northwestern Sichuan area, the province will build zero-carbon computing parks in places like Xinduqiao (Ganzi), Hongyuan (Aba), and Yanyuan (Liangshan). Many existing parks, including Ya'an Big Data Industrial Park, will also be upgraded. As computing power is key to development, Sichuan plans to build a high-level computing center, attracting major telecom operators and leading model companies to deploy 10,000-card-level intelligent computing clusters in the core areas and development belt, with individual project support of up to 10 million yuan.
To realize the value of computing power, the policy emphasizes universal and affordable access. The province will continuously upgrade its integrated computing power monitoring and scheduling platform for unified management and dispatch. It will also improve the application subsidy mechanism by introducing "computing vouchers" and "token vouchers," offering up to 1 million yuan in subsidies for purchasing domestic computing power and token consumption.
Creating benchmark application scenarios and reducing electricity costs, the policy targets the "AI+" sector. It supports building high-quality computing supply, datasets, and benchmark scenarios in areas such as industrial manufacturing, cultural tourism, and smart healthcare. The provincial budget will cover up to 25% of the total investment for datasets and AI corpora, with a maximum of 10 million yuan per project. It also encourages various entities to open application scenarios, supporting up to 30 million yuan for each key application scenario project in smart manufacturing, smart tourism, smart healthcare, disaster prevention, the low-altitude economy, and urban governance.
Electricity is critical for sustainable computing power and cost reduction. The policy supports data centers in participating in electricity market transactions, including energy, auxiliary services, and demand response. This allows them to adjust their electricity usage based on price signals, reducing costs. Local governments are encouraged to prioritize surplus electricity for computing companies. Projects with confirmed investment or construction by the end of 2027 will be exempt from high-reliability power supply fees. Finally, computing companies are encouraged to sign long-term green power purchase agreements with renewable energy generators, expanding green electricity and certificate trading to ensure a green consumption record.