China’s Ministry of Industry and Information Technology reported that by the end of 2025 the country’s three state‑owned carriers — China Mobile, China Telecom and China Unicom — were offering 938,000 data‑centre racks to external customers, an increase of 108,000 racks from the previous year. The ministry framed the growth as part of a deliberate strategic shift: carriers are moving their算力 (compute‑power) deployments from broad geographic coverage to deeper integration with networks and intelligence.
That shift, described in the ministry’s annual communications statistics, emphasises算网融合 — the fusion of computing and networking — and the construction of unified resource orchestration and intelligent scheduling capabilities. In practice, carriers say they are evolving from suppliers of basic cloud infrastructure toward providers of intelligent, green and diversified compute services, able to support latency‑sensitive applications, edge AI and large‑scale model workloads.
Rack counts are a blunt but widely used gauge of data‑centre capacity: they reflect the physical footprint carriers make available to customers, hosts and their own cloud platforms. The addition of more than 100,000 racks in a single year signals continued heavy investment in physical infrastructure at a time when demand for AI training, inference at the edge and industrial digitalisation is rising across China’s economy.
Commercially, the development marks the telcos’ attempt to climb the value chain. Historically focused on connectivity, the three carriers are leveraging extensive fibre, site access and customer relationships to offer compute as a service, competing with hyperscale cloud providers and third‑party data‑centre operators in domestic markets. Their push to integrate network intelligence with compute orchestration could lower latencies for enterprise AI deployments and create bundled offerings that are hard for pure cloud players to replicate.
The move also has material implications for energy use and sustainability. Carriers emphasise “green” compute and smarter scheduling to improve utilisation, but rapid rack additions increase power and cooling demands, creating pressure on regional grids and accelerating demand for renewable energy and efficiency technologies. How quickly carriers can raise utilisation rates and deploy energy‑saving innovations will shape the economics of their compute push.
Strategically, the telcos’ expanding compute footprint supports Beijing’s broader aims of technological self‑reliance and domestic control over critical digital infrastructure. A larger, more integrated carrier compute platform reduces dependence on foreign vendors for some services and creates a domestic ecosystem for AI, chips and cloud software. That could strengthen China’s internal supply chains while reshaping market dynamics for local cloud firms, hardware suppliers and enterprise customers.
The headline figure — 938,000 external racks — is less a one‑off statistic than a signpost. It shows the three carriers are not merely scaling capacity but changing how they position themselves in the digital economy: from pipes to platform providers offering managed, intelligent compute. Observers should watch the carriers’ product rollout, pricing behaviour and partnerships with chipmakers and software firms to gauge whether this infrastructural growth translates into sustainable, profitable services or merely a wave of capacity that industry will need to fill.