DEPLOYDatabase

Company

Alibaba Cloud

Alibaba Cloud (Aliyun) is the cloud and AI infrastructure arm of Alibaba Group, founded in September 2009 in Hangzhou and today the dominant public-cloud…

Founded
2009
HQ
Hangzhou, China (international HQ Singapore)
Status
active

Overview

Alibaba Cloud (Aliyun) is the cloud and AI infrastructure arm of Alibaba Group, founded in September 2009 in Hangzhou and today the dominant public-cloud operator in China and the Asia-Pacific market by Gartner and IDC share. Unlike the pure-play neocloud pack (CoreWeave, Nebius, Crusoe), Alibaba Cloud is a full hyperscaler: it sells IaaS, PaaS, databases, and CDN alongside GPU capacity, and its AI story rides on two proprietary pillars, the PAI (Platform for AI) training and inference stack and the Qwen family of open-weights foundation models that Alibaba itself trains, serves, and licenses.\n\nIn February 2025 the company said it would spend more than 380 billion yuan (about 53 billion US dollars) on cloud and AI infrastructure over the following three years, a figure larger than its combined AI-infra spend across the prior decade and the biggest disclosed AI capex commitment by any Chinese company.

The build-out is constrained on the silicon side by US export controls: Alibaba Cloud has publicly bought Nvidia's China-only H20 accelerators, has ramped in-house T-Head Yitian ARM CPUs, and is one of the biggest customers for domestic Chinese AI chips including Huawei Ascend, though it has not disclosed a total GPU count. It operates data centers across 29 regions and 87 availability zones globally as of its most recent disclosure, with 2025-2026 expansion pushes into Southeast Asia (Malaysia, Thailand, the Philippines), Mexico, and the Middle East (Dubai, Riyadh).\n\nWhat separates Alibaba Cloud from the Western neocloud set is that its anchor tenant is itself: Qwen inference, Taobao and Tmall commerce, Ele.me delivery, Cainiao logistics, and DingTalk collaboration all run on it, and the Qwen-3 series (culminating in Qwen3-Max at trillion-plus parameters, released in 2026) has become one of the most widely deployed open-weights model families outside the US.

Positioning is genuinely dual: it is a training platform for its own frontier models and for enterprise fine-tunes, and a very high-volume inference operator for Chinese and Southeast Asian consumer traffic. Alibaba Group trades on NYSE (BABA) and HKEX (9988); the cloud unit sits inside the Alibaba Group holding structure after a planned 2023-2024 spinout was called off.

Verified record

Verified deployments
None on file
Active incidents
None on file

Key facts

Founded

September 2009 as Aliyun, in Hangzhou, China

Parent + ticker

Wholly owned by Alibaba Group Holding (NYSE: BABA; HKEX: 9988); planned 2023 spinout of the cloud unit was cancelled in November 2023

Announced AI + cloud capex

More than 380 billion yuan (about 53 billion USD) committed to AI and cloud infrastructure over three years, announced February 2025

Global footprint

Data centers across 29 regions and 87 availability zones worldwide per company disclosure; the largest cloud operator in China and APAC per Gartner

GPU inventory

Not disclosed. Public reporting has Alibaba Cloud as one of the largest buyers of Nvidia's China-only H20 accelerator and of Huawei Ascend 910B/910C; no vendor-mix number has been released by the company

Announced hourly GPU pricing

Not published as a single global rate card; A100/H20 instances are quoted per-region on the Alibaba Cloud ECS console rather than in a headline reserved-vs-on-demand table

Silicon partnerships

Nvidia (H20 for China; H100/H200 for international regions), Huawei Ascend, AMD, plus in-house T-Head Yitian 710 ARM CPU launched October 2021

Anchor tenant

Itself: Qwen model training and inference, Taobao/Tmall commerce, DingTalk, Cainiao logistics, and Ele.me all run on Alibaba Cloud; external anchor customers include Xiaohongshu and much of the Chinese enterprise SaaS ecosystem

Proprietary model family

Qwen open-weights LLMs (launched April 2023); Qwen3-Max at roughly 2.4 trillion parameters released August 2026, with distilled variants under Apache 2.0

AI platform product

PAI (Platform for AI) is the training + inference layer; Qwen models are served through PAI-EAS and Model Studio

2025-2026 expansion

New availability zones opened in Malaysia, the Philippines, Thailand, Mexico, and the UAE as part of the three-year capex commitment

Positioning

Full hyperscaler doing both training and inference at hyperscale volumes; unlike Western neoclouds it is its own largest AI customer via Qwen and Alibaba consumer properties

Data & sources

News coverage

1

Web sources

4

5 sources backing this record.View all →

Alibaba Cloud on the deployment map

View the global map

Alibaba Cloud has no verified deployments on the DEPLOY registry yet. Explore the deployment map by place and type.

Free to use on your site, with a link back to DEPLOY.

No verified deployments for Alibaba Cloud on the DEPLOY record yet. Track Alibaba Cloud to know the moment the first one is verified.

Current leadership (3)

  • Eddie Wu CEO, Alibaba Group; direct oversight of Alibaba Cloud Intelligence Group since September 2023reported, not verified
  • Joe Tsai Chairman, Alibaba Groupreported, not verified
  • Jingren Zhou CTO, Alibaba Cloud Intelligencereported, not verified

Safety record

No incidents on record for Alibaba Cloud.

Only active incidents are counted. Retracted incidents are excluded from this summary but remain reachable at their canonical URLs.

Full safety record: incidents, sourcing, and exposure data →

Alibaba Cloud on the data-center map

Every verified data center Alibaba Cloud operates that has a mapped coordinate (9 total). Filled circles are Tier-2 facilities and Tier-1 campuses under this operator; click any pin to open its record.

Alibaba Cloud data centers (15)

AI data centers Alibaba Cloud operates or is a party to, from the DEPLOY data center registry.

Recent coverage

Alibaba Cloud in third-party press