Alibaba Cloud: China's AI Cloud and Qwen Models | hokai.io
Alibaba Cloud, founded 2009, is China's no.1 AI cloud with 35% market share and Qwen AI models. Q1 2026 cloud revenue up 26% to $4.7B, triple-digit AI growth.
Alibaba Cloud is the cloud and AI division of Alibaba Group (NYSE: BABA), founded in 2009 and headquartered in Hangzhou, China. It leads China's cloud market with 35% share, operates 94 availability zones in 29 regions, and committed $53B in AI and cloud infrastructure through 2027. The Qwen AI model family tops 1 billion Hugging Face downloads with the consumer app reaching 300 million monthly users by February 2026.
Alibaba Cloud, founded in 2009 and headquartered in Hangzhou, China, is the cloud and AI division of Alibaba Group (NYSE: BABA). It holds 35% of China's cloud market and 24.6% of China's AI public cloud. The Qwen AI model family surpassed 1 billion Hugging Face downloads in January 2026. Q1 2026 cloud revenue reached $4.66 billion, up 26% year-over-year, with AI revenue growing at triple-digit rates.
Founded: 2009 · HQ: Hangzhou, Zhejiang, China · Team: 4000-5000 · CEO: Eddie Wu · Funding: Division of Alibaba Group (NYSE: BABA). FY2025 Group revenue $137.3B. Cloud Intelligence Group Q1 FY2026 revenue $4.66B (+26% YoY). Group committed $53B in AI/cloud capex 2025-2027.
About Alibaba Cloud
Alibaba Cloud, known in Chinese as Aliyun (阿里云), was founded in 2009 as the cloud computing arm of Alibaba Group Holding Limited, itself co-founded by Jack Ma in Hangzhou, Zhejiang, China, in 1999. The cloud division was established to provide Alibaba's e-commerce platform with scalable infrastructure before expanding to serve external customers globally across more than 200 countries and regions. Alibaba Group is publicly traded on NYSE as BABA and Hong Kong Stock Exchange as ticker 9988. Under CEO Eddie Yongming Wu, who assumed the Group CEO role in September 2023, Alibaba Cloud has become the company's most critical growth engine, receiving the majority of a committed $53 billion three-year AI and cloud capital expenditure program. Alibaba Cloud operates across three main product families: general cloud infrastructure, the Qwen AI model family, and enterprise AI platforms. General cloud infrastructure includes Elastic Compute Service (ECS) for virtual machines, Object Storage Service (OSS), ApsaraDB for relational and NoSQL databases, MaxCompute for big data analytics, and Virtual Private Cloud (VPC) networking. The Qwen (Tongyi Qianwen) model family is the flagship AI offering, spanning text generation, vision, audio, image creation, and agentic workflow capabilities across both open-source and proprietary tiers. Enterprise platforms include Model Studio (API access to Qwen models with OpenAI-compatible and Anthropic-compatible interfaces), PAI (Platform for AI for training and deployment), and Wukong (an enterprise AI agent platform for complex business task automation). In April 2026, Alibaba released Qwen3.6-Plus, featuring a 1M-token context window and markedly improved agentic coding capability, available via the Alibaba Cloud Bailian platform and API. Qwen3.5, released February 16, 2026, includes a 397B-A17B mixture-of-experts model described as the third-best open-source model globally, benchmarked as comparable to GPT-5 and Claude Opus 4.5. Also in April 2026, Alibaba formed the Alibaba Token Hub (ATH), a new AI business unit led by CEO Eddie Wu that consolidates the Tongyi large model division, Qwen, Wukong, and MaaS operations under a single mandate. The Qwen consumer app reached 300 million monthly active users by February 2026, and the Qwen model family surpassed 1 billion cumulative downloads on Hugging Face on January 21, 2026. Alibaba Group committed RMB 380 billion (approximately $53 billion) over three years in AI and cloud infrastructure, with reports as of late 2025 suggesting the figure may increase to $69 billion. The company spent 120 billion yuan ($16.8 billion) on AI and cloud capital expenditure over the four quarters through end-2025. Cloud Intelligence Group Q1 FY2026 (June 2025) revenue reached $4.66 billion, up 26% year-over-year, while Q2 FY2026 (September 2025) cloud revenue of 39.8 billion yuan grew 34% year-over-year, the fastest rate in two years. AI-related cloud revenue has delivered triple-digit year-over-year growth for more than 10 consecutive quarters as of May 2026. Alibaba Cloud earns revenue through pay-as-you-go consumption of cloud infrastructure, API model inference through Model Studio, enterprise AI platform subscriptions, and managed AI solutions built on Qwen and PAI. The company holds approximately 35% of China's overall cloud market and 24.6% of China's AI public cloud market, ranking first in both categories for the second consecutive year per IDC's August 2025 data. Token consumption on Alibaba's model platform increased sixfold in just three months in early 2026, reflecting rapid enterprise adoption. Alibaba Group's total fiscal year 2025 revenue reached $137.3 billion, with the cloud division as the primary growth driver and profit margin expander. Eddie Yongming Wu, Alibaba Group CEO since September 2023 and one of Alibaba's original co-founders, directly leads the Alibaba Token Hub AI initiative. Joseph C. Tsai is Alibaba Group Chairman. Li Feifei became Alibaba Cloud CTO in April 2026; former Cloud CTO Zhou Jingren, who also led the Tongyi AI lab and joined the Alibaba Partnership in October 2025, now serves as Chief AI Architect on the newly formed Group Technology Committee. Alibaba Cloud as a standalone entity employs approximately 4,000-5,000 people per PitchBook and RocketReach data; Alibaba Group overall employed approximately 128,000 people as of early 2026, a 34% reduction from prior years following major retail divestitures. Alibaba Cloud's mission is to provide reliable AI and cloud computing services to enterprises globally, with open-source model leadership as a core pillar. The Qwen family has released over 100 open-weight models, with over 200,000 derivative models now on Hugging Face's platform. Qwen3, released April 28, 2025, was trained on 36 trillion tokens across 119 languages and dialects, advancing benchmarks in coding, math, multilingual reasoning, and tool use. The company has also received Chinese government approval to purchase up to 400,000 NVIDIA H200 GPUs (alongside ByteDance and Tencent), and has integrated the full NVIDIA Physical AI software stack into PAI for humanoid robotics development. Alibaba Cloud leads China's cloud market against domestic rivals Huawei Cloud, Tencent Cloud, and Baidu AI Cloud. Globally, it competes with AWS, Microsoft Azure, and Google Cloud, where its footprint is smaller but it holds a sovereign compliance advantage in China-specific regulatory environments. The Qwen model family competes with Meta Llama, Mistral AI, and Chinese rivals DeepSeek and Baidu ERNIE; Qwen3.5-397B ranked as the third-best open-source model as of February 2026 on major coding and reasoning benchmarks. Against AWS and Azure in global enterprise cloud, Alibaba Cloud wins on China data sovereignty and local government trust but faces significant disadvantages in brand recognition and established enterprise relationships outside Asia. Alibaba Cloud operates under Chinese government data regulations, US export control restrictions on advanced Nvidia chips, and EU GDPR requirements for its international data centers in Europe. The Pentagon added Alibaba to its list of Chinese military companies in February 2026, creating market volatility for BABA shares despite carrying no direct operational restrictions. US chip export curbs constrain access to the most advanced NVIDIA GPUs, though the Chinese government approved a controlled H200 GPU allotment shared among Alibaba, ByteDance, and Tencent. The cloud spinoff originally planned under Alibaba's 2023 restructuring was reversed due to geopolitical risks. EU operations will be anchored by new data centers launching in France and the Netherlands during 2026. Near-term priorities include the commercial rollout of Wukong and Alibaba Token Hub, international data center openings in Brazil, France, the Netherlands, Mexico, Japan, and South Korea, and continued Qwen model releases under both open Apache 2.0 and proprietary licenses. CEO Eddie Wu has publicly targeted $100 billion in combined cloud and AI annual revenue within five years, driven by enterprise Qwen adoption and model-as-a-service growth. The $53 billion infrastructure commitment positions Alibaba Cloud to compete directly with AWS and Azure in Asia-Pacific emerging markets while maintaining its dominant position in Greater China.
Mission
Enable enterprises globally to access reliable, scalable AI and cloud computing services that drive innovation and business growth.
Products
- Qwen AI Models (Open-source and proprietary LLM and multimodal model family): https://www.alibabacloud.com/en/solutions/generative-ai/qwen
- Model Studio (API platform for Qwen model inference and fine-tuning): https://www.alibabacloud.com/help/en/model-studio
- Alibaba Cloud Infrastructure (Cloud compute, storage, and database services (ECS, OSS, ApsaraDB)): https://www.alibabacloud.com/en/product
- PAI (Platform for AI) (Managed ML training, fine-tuning, and deployment platform): https://www.alibabacloud.com/en/product/machine-learning
- Wukong (Enterprise AI agent platform for business task automation): https://www.alibabacloud.com
Compliance
SOC 2 Type II, SOC 3, ISO 27001, ISO 27017, ISO 27018, ISO 27701, ISO 22301, PCI DSS, HIPAA-eligible, GDPR, CSA STAR, C5 (Germany), MTCS (Singapore)
Links
Website · GitHub · Twitter · LinkedIn · Blog · Docs
Frequently Asked Questions
What is Alibaba Cloud and what do they build?
Alibaba Cloud, also known as Aliyun (阿里云), is the cloud computing and AI division of Alibaba Group Holding Limited (NYSE: BABA), founded in 2009 and headquartered in Hangzhou, Zhejiang, China. The division builds and operates cloud infrastructure, AI foundation models, and enterprise AI platforms for customers in more than 200 countries and regions. Core cloud offerings include Elastic Compute Service (ECS), Object Storage Service (OSS), ApsaraDB for managed databases, and MaxCompute for big data analytics. The Qwen (Tongyi Qianwen) AI model family covers text, vision, audio, image generation, and agentic reasoning, available in both open-source (Apache 2.0) and proprietary API tiers through Model Studio. Enterprise platforms include PAI (Platform for AI) for model training and deployment, and Wukong for enterprise AI agent automation of complex business workflows. As of February 2026, the Qwen consumer app has reached 300 million monthly active users, and the Qwen model family surpassed 1 billion cumulative Hugging Face downloads in January 2026.
Who founded Alibaba Cloud and who is the CEO?
Alibaba Cloud was established in 2009 as an internal cloud computing initiative within Alibaba Group, which was co-founded by Jack Ma and 17 partners in Hangzhou, China, in 1999. The cloud division was created to provide Alibaba's growing e-commerce operations with scalable infrastructure, with Wang Jian (known as Dr. Jian) widely credited as its technical architect and first CTO. Alibaba Group CEO Eddie Yongming Wu, one of the original Alibaba co-founders and former chairman of Taobao and Tmall Group, has served as the Group's CEO since September 2023 and directly oversees the Alibaba Token Hub AI initiative launched in April 2026. Alibaba Group Chairman is Joseph C. Tsai, co-founder and former executive vice chairman of Alibaba. In April 2026, Li Feifei was appointed Alibaba Cloud CTO; former Cloud CTO Zhou Jingren, who also led the Tongyi AI lab, joined the Alibaba Partnership in October 2025 and now serves as Chief AI Architect on the Group Technology Committee. Jack Ma, who served as Executive Chairman until September 2019, relocated to Tokyo in 2022 and has largely stepped back from public business activity.
How much revenue does Alibaba Cloud generate?
Alibaba Cloud Intelligence Group Q1 FY2026 (quarter ending June 2025) revenue reached $4.66 billion (33.9 billion yuan), up 26% year-over-year. Q2 FY2026 (September 2025 quarter) cloud revenue of approximately 39.8 billion yuan grew 34% year-over-year, the fastest growth rate in more than two years, driven by AI workloads. AI-related cloud revenue has maintained triple-digit year-over-year growth for more than 10 consecutive quarters as of May 2026. Cloud division EBITA grew 35% to 3.6 billion yuan in Q2 FY2026, confirming that AI demand is translating to margin expansion and not just top-line growth. Alibaba Group's full fiscal year 2025 (ending March 2025) total revenue was $137.3 billion, with cloud as the primary growth segment. The company has committed $53 billion in AI and cloud capital expenditure through 2027, with reports suggesting this may increase to $69 billion as AI demand accelerates.
What products does Alibaba Cloud make?
Alibaba Cloud's product portfolio spans general cloud infrastructure, AI model services, and enterprise AI platforms. Infrastructure products include Elastic Compute Service (ECS) for virtual machines across hundreds of instance types, Object Storage Service (OSS), ApsaraDB for relational databases and RDS, Virtual Private Cloud (VPC) networking, and MaxCompute for big data analytics. The Qwen AI model family is the flagship AI offering, spanning Qwen3 text models (released April 2025, trained on 36T tokens across 119 languages), Qwen-VL for vision-language tasks, Qwen-Audio and Qwen3-TTS for audio, Qwen-Image-2512 for open-source image generation, and Qwen3.6-Plus (April 2026) for agentic coding with a 1M-token context. Model Studio provides OpenAI-compatible and Anthropic-compatible API interfaces for Qwen model inference and fine-tuning. Wukong is an enterprise AI agent platform for automating complex business tasks using multiple AI agents and corporate data integration. PAI (Platform for AI) handles managed model training, fine-tuning, and deployment with NVIDIA GPU and Alibaba's own Hanguang AI chip support.
Where is Alibaba Cloud headquartered and how big is the team?
Alibaba Cloud is headquartered in Hangzhou, Zhejiang, China, alongside Alibaba Group's corporate campus; it also maintains major engineering and sales offices in Beijing, Shanghai, and Singapore. As a standalone entity, Alibaba Cloud employs approximately 4,000-5,000 people per PitchBook (4,139) and RocketReach (4,921) data; these figures represent the cloud computing unit's direct staff, not Alibaba Group's broader workforce. Alibaba Group overall employed approximately 128,197 people as of December 2025, down from 194,320 a year earlier, a 34% reduction driven by the sale of Sun Art retail and other non-core asset divestitures. The company operates 94 availability zones across 29 global regions, with data centers in China, Singapore, Japan, Australia, Malaysia, Indonesia, India, Germany, the UK, the US, UAE, and more. In 2025, Alibaba Cloud expanded into Mexico (first Latin American data center) and launched its second South Korea and third Malaysia data centers. New data centers are launching in Brazil, France, and the Netherlands in 2026.
What is Alibaba Cloud's mission or research focus?
Alibaba Cloud's stated mission is to enable enterprises globally to access reliable AI and cloud computing that drives innovation, with open-source model leadership as a strategic differentiator. The Qwen research team has released over 100 open-weight models under Apache 2.0 licensing, generating over 200,000 derivative models on Hugging Face and surpassing 1 billion cumulative downloads in January 2026. Qwen3, released April 2025, was trained on 36 trillion tokens across 119 languages and dialects, supporting hybrid thinking modes (standard and extended reasoning) for flexible cost-performance tradeoffs. Research areas include mixture-of-experts architectures (Qwen3.5-397B-A17B activates only 17B of 397B parameters during inference), multimodal reasoning combining text, vision, and audio, and agentic AI for autonomous task execution in enterprise environments. CEO Eddie Wu created the Alibaba Token Hub in April 2026 to unify AI model development, inference infrastructure, and enterprise AI products under one strategic unit. Alibaba Cloud is also collaborating with NVIDIA on Physical AI software for humanoid robotics and has integrated NVIDIA's full physical AI stack into its PAI platform.
Is Alibaba Cloud compliant with SOC 2, GDPR, ISO 27001, and HIPAA?
Alibaba Cloud holds an extensive compliance certification portfolio, verified by independent third-party auditors and documented at alibabacloud.com/trust-center/compliance. Certifications include SOC 2 Type II, SOC 3, ISO 27001, ISO 27017 (cloud security controls), ISO 27018 (cloud privacy), ISO 27701 (privacy information management), ISO 22301 (business continuity), PCI DSS, HIPAA-eligible, and CSA STAR. Regional certifications include Germany's C5 standard, Singapore's MTCS (Multi-Tier Cloud Security), TISAX for the automotive industry, Australia's APRA for financial services, and Hong Kong's HKMA and SFC frameworks. For GDPR compliance, Alibaba Cloud provides a Data Processing Addendum incorporating Standard Contractual Clauses under GDPR Article 46, and acts as a data processor for customer personal data, with customers retaining the controller role. The shared responsibility model means Alibaba Cloud secures its infrastructure layer while customers configure security controls and data access for their own applications. New EU data centers in France and Netherlands launching in 2026 will allow EU customers to keep data fully within the European Economic Area.
Who are Alibaba Cloud's main competitors?
Alibaba Cloud competes domestically in China against Huawei Cloud, Tencent Cloud, and Baidu AI Cloud, where it holds approximately 35% overall market share and 24.6% of AI public cloud per IDC's August 2025 data. Against Huawei Cloud, Alibaba wins on broader international presence and open-source Qwen brand recognition; Huawei wins on government hardware procurement and telecom integration where Alibaba lacks relationships. Against Tencent Cloud, Alibaba wins on AI model sophistication and developer ecosystem depth; Tencent wins on gaming and social platform integration. Globally, Alibaba Cloud competes with AWS, Microsoft Azure, and Google Cloud, where it is significantly smaller by revenue and enterprise relationships but holds regulatory advantages for China-based workloads neither AWS nor Azure can serve. In the open-source AI model space, Qwen competes with Meta's Llama family and Mistral AI; Qwen3.5-397B ranked as the third-best open-source model globally as of February 2026. Chinese AI rivals DeepSeek and Baidu ERNIE represent the most direct domestic model competition, though Qwen's download volume significantly exceeds both.