For the first half of 2026, China Mobile reported RMB52.9B of computing-services revenue, up 14.0%; RMB19.0B of data-center revenue, up 13.2%; AIDC revenue growth of 486.1%; and RMB5.3B of intelligent-computing-services revenue, up 130.1%.
The headline percentage should not be detached from base and delivery. The release does not give a stand-alone AIDC revenue amount, nor does it translate the revenue directly into GPUs, MW or project capacity.
That makes the results a useful case study of how a carrier embeds AI infrastructure in a service mix, not a basis for extrapolating all China AIDC installed capacity.
Four compute measures in one interim report
Computing services and data-center revenue cover broader service sets; AIDC and intelligent-computing services are more direct AI-commercialization signals. Their overlap and internal composition are not separately detailed, so they must not be mechanically added.
China Mobile also reported a 4+N+31+X resource footprint, 112,700 PFLOPS of FP16 intelligent-computing scale and network-cloud resource utilization above 90%. These are resource and operating-efficiency measures, not equivalent to sold or recognized AI revenue.
| Metric | Disclosure | Basis and boundary |
|---|---|---|
| Computing services | RMB52.9B | H1 2026, +14.0% year on year. |
| Data-center revenue | RMB19.0B | H1 2026, +13.2% year on year. |
| Intelligent compute services | RMB5.3B | H1 2026, +130.1% year on year. |
| AIDC revenue | +486.1% | Company-reported growth; no stand-alone amount disclosed. |
The carrier moat is connection and delivery, not racks alone
A telecom operator can place network, cloud, IDC, industry customers and operations inside one sale-and-delivery chain. For enterprise buyers, connectivity, data governance, deployment and long-term service can matter as much as GPU supply.
That integration also makes revenue harder to interpret: attribution across network, cloud and data center, capital intensity and utilization decide whether rapid growth becomes durable cash return.
Access and transport
Enterprise data, dedicated links and low-latency transport connect model services to operating workflows.
Platform and scheduling
Multi-tenancy, industry software and scheduling decide whether compute becomes a reusable service.
Delivery and operations
Power, halls, liquid cooling and field operations determine whether resources come online as promised.
The next test is revenue quality, not one-quarter growth
Watch for AIDC absolute revenue, customer mix, contract terms, utilization, margin and capex efficiency. Only together can they show whether AI services are one-off projects or a repeatable operating platform.
This article relies on China Mobile's primary disclosure and does not merge undeclared peer figures or media leads into the comparison.
IDC ATLAS VIEWAIDC growth merits attention, but its value must be confirmed by utilization, customer stickiness and cash conversion. A carrier's edge is putting compute inside a deliverable network service, not merely placing servers in a hall.