China puts almost 490 billion euros into networks and computing, targeting 9,800 Eflops

China published its 15th five-year plan for information and communications on 8 September 2026. The plan covers 2026-2030 and puts cumulative investment at around 490 billion euros. It sets the direction for public spending and incentives on networks, computing and cybersecurity over the next five years.

The targets

The plan aims to reach 9,800 Eflops of intelligent computing capacity by 2030. It breaks the goal into thirteen indicators and six lines of action. Sector revenue is expected to come in at around 525 billion euros. The infrastructure goals include computing clusters with at least 100,000 accelerators, a threshold that puts individual sites in the class of large hyperscale facilities. The document also calls for strengthening 5G Advanced, research into 6G, ultra-broadband, and work on cybersecurity and data protection. It places these initiatives within programmes that coordinate eastern and western regions.

On the ground: Wuhan's 6G hub

Wuhan shows how the plan turns into building sites. In the Future Sci-Tech City, the 光谷 area, a test zone for the 6G industry has been launched. It involves China Mobile Hubei, CITIC Mobile, the Zhongguancun Panlian Institute and the Huazhong University of Science and Technology. By 2027 the programme foresees three pieces of infrastructure: the country's largest integrated 6G experimental network, an open device for integrating communication and sensing, and an open platform for intelligent metasurfaces. The test stations cover about 3.3 square kilometres.

The European reading

For Europe the plan is a distance gauge. The Union has concentrated its efforts on AI Gigafactories and on regulation. Beijing adds the research phase on next-generation networks to the industrial scale of computing. The document cites territorial coordination measures that shift computing loads towards regions with greater energy availability, and a programme integrating artificial intelligence with communication infrastructure.

The point is not so much competition over the number of installed accelerators as the ability to connect them. Without low-latency networks and without energy available at contained costs, the computing power declared in plan documents remains partly theoretical.