Alibaba
Chinese e-commerce and cloud group that develops the Qwen AI models
Profile
- Country
- China
- Headquarters
- Hangzhou, Zhejiang
- Founded
- 1999
- Founders
- 18 people led by Jack Ma
- Type
- Listed holding company (Cayman Islands; NYSE: BABA, HKEX: 9988)
Alibaba Group is a company headquartered in Hangzhou, Zhejiang Province, China. It was founded in 1999 by 18 people led by Jack Ma, a former English teacher. Legally, it is a holding company (a company that owns other companies) registered in the Cayman Islands, and it is listed on both the New York Stock Exchange (ticker BABA) and the Hong Kong Stock Exchange (stock code 9988). Some of its businesses inside China are run through a ‘VIE’ structure, in which it controls local companies through contracts rather than shareholdings. As of October 2026, its chairman is Joe Tsai and its CEO is Eddie Wu.
Alibaba's main businesses are online shopping in China (Taobao and Tmall), international online shopping (such as AliExpress) and cloud computing (Alibaba Cloud). Most of its revenue in the 2026 financial year came from these three business groups. The company says Alibaba Cloud is shifting the focus of its growth from renting out computing and storage towards selling AI models, AI computing power and agent services. It also has other businesses, such as logistics (Cainiao) and a maps app (Amap).
Qwen is a family of large language models and multimodal models (models that handle text, images and more together) made by Alibaba. In March 2026, Alibaba created the Alibaba Token Hub (ATH) Business Group, led directly by its CEO, bringing together Tongyi Lab, which researches foundation models, the MaaS business, which offers models as a service, the Qwen app business unit and other units. The Qwen app is a consumer AI assistant linked to services such as Taobao and Alipay. Several Qwen models are published on Hugging Face, but the licence (terms of use) differs from model to model. Qwen3.8-27B uses ‘Apache 2.0’, while Qwen3.8-2.4T-A95B, with 2.4 trillion parameters (the numbers a model adjusts as it learns), uses the ‘Qwen3.8-Max License’. Under that licence, a company that runs a business letting others use AI models over the internet (MaaS) or a coding or office AI assistant, and whose revenue together with its affiliates exceeds US$50 million over twelve months, must obtain a separate licence before using the model for any commercial purpose.
Sources for this profile (8) · as of Oct 22, 2026
- Introduction to Alibaba Group · Alibaba Group
- FAQs – Corporate Information · Alibaba Group
- Alibaba Group Holding Ltd – Form 20-F for the fiscal year ended March 31, 2026 · U.S. Securities and Exchange Commission (EDGAR)
- Alibaba Group Holding Limited Fiscal Year 2026 Annual Report (Form 6-K, Exhibit 99.1) · U.S. Securities and Exchange Commission (EDGAR)
- Alibaba Establishes Alibaba Token Hub Business Group · Alizila (Alibaba Group news site)
- Alibaba Group Holding Limited – announcement with Form S-8 (board composition as at 14 Oct 2026) · Alibaba Group Holding Limited (HKEXnews)
- Qwen3.8-27B model card · Hugging Face (Qwen)
- Qwen3.8-2.4T-A95B model card and LICENSE (Qwen3.8-Max License) · Hugging Face (Qwen)
AIMade by AIThis profile was written by AI (Claude) from the 8 sources above (official pages, filings, news reports).How it's made
Records 17
Newest first. Open one to see it with its source.
2026 8
A case of a frontier open-weight model disclosing the rationale for its architecture choices ablation by ablation. In the authors' own reporting, loss and downstream accuracy do not always move together, and combining th
China's largest AI infrastructure fundraising. A race to secure compute that accepts equity dilution
A practice in which the promise to open up a top model and its actual release come apart. It was pointed out at the same time that Kimi K3 had done the same — grounds for recording the announcement date and the weight re
While the top model moves to its own license, the small model stays on Apache 2.0 — a two-track licensing split by scale
The top open-weight model reverts to a custom license rather than Apache. Qwen3.8-Max is a closed commercial version built on it
A dense 27B beat its 397B MoE predecessor, and the center of gravity of truly open models moved down to the 30B class
Open-weight omni-modal models caught up with the frontier, handling 10 hours of audio and 400 seconds of video
Applies early-fusion multimodal training to a large Apache 2.0 MoE. Plus remains closed, so open and closed run in parallel
2025 4
Alibaba claimed state-of-the-art results among open models in agentic coding, browser use and tool use, and open-sourced the command-line coding tool Qwen Code alongside it
The point where open-weight embeddings overtook commercial API models on the leaderboard
A strategy of covering every size class with an unrestricted license overtook Llama in the number of derivative models on HF
The full release following QwQ-32B-Preview; Alibaba claims performance comparable to DeepSeek-R1 (671B total, 37B active) at 32B
2024 3
The first open reasoning model after o1
All open models except the 3B and 72B were released under Apache 2.0, and the 14B and 32B sizes missing from the previous generation returned
Only the 72B kept the original Qianwen License while the other four moved to Apache 2.0, the release where Qwen's licensing began to open up
2023 2
A sign that small open-weight models were starting to overtake commercial APIs
Alibaba's first open-weight language model; the original release was trained on 2.2T tokens, raised to 2.4T in an update on September 25, 2023