CHINA DATA CENTERS · PROJECT PROFILE

Huawei Cloud Gui’an: Linking Facility Efficiency with AI Services

Huawei’s Gui’an disclosures connect cooling and power engineering with the delivery of AI cloud services.

Gui’an, Guizhou, ChinaSources reviewed 23 October 2026
Huawei’s photograph of the Gui’an cloud campus, published with its account of data center efficiency.
Huawei’s photograph of the Gui’an cloud campus, published with its account of data center efficiency. Photo: Huawei.

At a glance

Operator: Huawei Cloud

Documented status: Operating cloud and AI computing location identified in corporate disclosures

What the sources establish

Huawei’s sustainability case study describes natural ventilation, heat reuse, liquid-cooling technology and changes to power-supply components at its Gui’an data center. It reports a PUE of 1.12 in that case. Separately, Huawei’s 2024 annual report identifies Gui’an among its Chinese AI computing centers, alongside Ulanqab and Wuhu. Huawei: Improving PUE to build green data centers Huawei Annual Report 2024: AI computing centers

These sources address different questions: the sustainability story describes facility engineering, while the annual report establishes the location’s role in the AI cloud platform.

Why the engineering boundary matters

Facility efficiency can support economical computing, but it does not specify how a model performs or how much usable AI capacity a customer receives. The connection between the two requires an agreed hardware configuration, software environment and workload test.

An international reader can use this example to separate three layers of evidence: the building’s energy systems, the computing platform and the delivered service. Heat-recovery benefits should also be evaluated against the actual receiving load and operating season, rather than treated as a constant reduction under every condition.

Evidence and comparison limits

The PUE of 1.12 is attributed to Huawei’s historical case study, not independently verified as a 2026 annual average. No current accelerator count or total IT MW is asserted.

What to follow

Follow newer facility measurements and workload-specific service evidence, especially the boundary between campus infrastructure and AI platform capacity.

Original information and official project sources

Open the underlying operator material, public-authority disclosure or certification register below. Source dates and reporting periods are separate from this profile’s publication date.

Huawei: Improving PUE to build green data centers ↗Source date / reporting period: Live story; publication date not shown · Reviewed 23 October 2026Huawei Annual Report 2024: AI computing centers ↗Source date / reporting period: 2024 reporting year · Reviewed 23 October 2026

← Back to China Data Centers