Conceptual diagram of AI Photo: VCG

The Ministry of Industry and Information Technology (MIIT) on Thursday launched a special action plan to boost computing power for small and medium-sized enterprises (SMEs).

The ministry said it proposed innovative services such as "computing power banks" and "computing power supermarkets" to support SMEs.

The move is in line with the 15th Five-Year Plan (2026-30) for National Economic and Social Development, which vowed to improve the inclusive and user-friendly access to computing power, reduce computing costs for SMEs and better support their digital transformation and intelligent development.

The aim is to guide basic telecommunications enterprises, computing power service providers, and other entities to build advanced storage capacity centers tailored to SMEs, providing nearby access to massive data supply resource pools, and to offer SMEs services such as high-performance storage and cross-region artificial intelligence (AI) data platforms, thereby strengthening storage capacity supply capabilities. 

It also encourages local authorities to provide subsidies for eligible SMEs, such as "computing power vouchers," while simplifying the processes for obtaining and using such vouchers.

Computing power banks and computing power supermarkets are cloud computing variants that treat computing power as a tradable resource. Computing power supermarkets enable transactions between suppliers and SME buyers, while computing power banks offer SMEs "use now, pay later" quotas and allow suppliers to secure collateralized loans, Pan Helin, a member of the experts committee of the MIIT, told the Global Times on Thursday.

The move aims to reduce computing costs for SMEs while improving the convenience of computing power purchasing channels, thereby encouraging SMEs to undergo digital and intelligent transformation, Pan said.

In Pan's view, SMEs currently have limited enthusiasm for digital and intelligent transformation. This is partly because they are unfamiliar with the potential integration of AI into their businesses, and partly due to the high cost of computing power.

The new measures could guide them toward a path of digital and intelligent development, and accelerate the implementation of the nation's "AI+" strategy, Pan added. 

By the end of 2028, a comprehensive, low-cost, high-service, vibrant ecosystem, and talent-strong, inclusive computing power service system shall be basically established, covering no fewer than 10 of the 15 industry categories applicable under the SME classification standards, according to the plan.

Public services for computing power applications by SMEs will be further strengthened, the barriers for SMEs to access and use computing power will be significantly lowered, and solid computing power support will be provided to promote the specialized and sophisticated development of SMEs that produce new and unique products, said the plan.

China had more than 6,000 AI enterprises last year, and the scale of the country's core AI industry is expected to have exceeded 1.2 trillion yuan ($171.39 billion), Xinhua News Agency reported in January, citing data from the MIIT.

In 2025, domestic companies released multiple AI chip products, and the country's intelligent computing power capacity reached 1,590 EFLOPS last year, said Zhang Yunming, a MIIT vice minister.

EFLOPS, or ExaFLOPS, refers to a unit measuring computing performance, with each unit representing 1 quintillion floating-point operations per second.