Key Highlights
- The Zhenwu M890 AI accelerator delivers triple the computational power of the previous 810E model, designed specifically for agentic AI applications.
- T-Head, Alibaba’s chip design division, created the processor now accessible to Chinese businesses through Alibaba Cloud services.
- The company revealed a strategic chip development timeline featuring the V900 (launching Q3 2027) and J900 (arriving Q3 2028), each targeting approximately 3x performance improvements.
- The tech giant simultaneously introduced Qwen 3.7-Max, an enhanced language model optimized for sophisticated programming tasks and extended agent operations lasting up to 35 hours.
- Alibaba has distributed over 560,000 Zhenwu processors to more than 400 clients spanning 20 different sectors.
On Wednesday, Alibaba introduced its latest internally-developed AI accelerator and an enhanced version of its primary large language model, reinforcing its commitment to establishing a self-sufficient AI infrastructure that doesn’t rely on Nvidia hardware.
The newly announced processor, designated Zhenwu M890, provides triple the computational capacity compared to the earlier Zhenwu 810E model. The reveal took place during the Alibaba Cloud Summit held in Hangzhou, with immediate availability for Chinese corporate clients via Alibaba Cloud’s Bailian infrastructure.
Shares of Alibaba (BABA) dipped 0.60% following the product announcements.
Alibaba Group Holding Limited, BABA
T-Head, Alibaba’s semiconductor engineering division, designed the M890. The chip features 144GB of GPU memory, representing a significant increase from the 96GB capacity in its predecessor, enabling substantially larger dataset processing.
The M890 marks a departure from the previous generation by supporting both model training and inference operations, whereas earlier versions handled inference exclusively. This dual capability positions it as a comprehensive solution for organizations developing and deploying AI-powered applications.
The accelerator specifically targets agentic AI implementations — autonomous systems capable of executing sophisticated, sequential operations with minimal human oversight. Such applications demand models that can maintain extensive contextual information and synchronize activities across multiple processes simultaneously.
According to T-Head, cumulative Zhenwu chip shipments have surpassed 560,000 units, serving more than 400 business clients across 20 sectors, including automobile manufacturing and financial services.
Extended Development Timeline Revealed
The company went beyond a single product launch, presenting an extended chip evolution strategy encompassing the V900, anticipated in Q3 2027, and the subsequent J900 scheduled for Q3 2028.
Projections indicate the V900 will achieve approximately another 3x performance enhancement compared to the M890. This consistent development rhythm demonstrates that Alibaba views semiconductor innovation as a strategic, ongoing infrastructure investment rather than an isolated initiative.
Complementing the M890 announcement, Alibaba introduced the Panjiu AL128, a server configuration that integrates 128 of the new accelerators within a single rack-mounted system.
Enhanced Language Model Released
Simultaneously, Alibaba launched Qwen 3.7-Max, the newest iteration of its leading language model. This version has been specifically optimized for complex programming challenges and extended autonomous agent operations, maintaining consistent performance during continuous operation spanning up to 35 hours.
Both the Qwen enhancement and the M890 chip were presented together at the Hangzhou conference, establishing them as interconnected components of a unified strategy focused on agentic AI capabilities.
This development arrives amid ongoing U.S. government restrictions limiting exports of Nvidia’s cutting-edge processors to Chinese entities. As a historically significant Nvidia client, Alibaba has accelerated its transition toward domestically-produced alternatives in response to these trade limitations.
Additional Chinese semiconductor firms such as Biren Technology and MetaX have recently completed public listings on mainland Chinese and Hong Kong stock markets, indicating widespread expansion throughout the industry.
Alibaba previously announced a commitment exceeding 380 billion yuan (approximately $53 billion) for cloud computing and AI infrastructure investments spanning three years — representing the company’s largest-ever financial commitment to this domain.
During a recent quarterly earnings discussion, CEO Eddie Wu stated that revenue generated from AI-related products is projected to emerge as the dominant growth engine within the cloud services division. The organization has characterized its proprietary chip solutions as delivering “the highest value for money compute power,” with potential to enhance profit margins.
Get 3 Free Stock Ebooks
Discover top-performing stocks in AI, Crypto, and Technology with expert analysis.
- Top 10 AI Stocks - Leading AI companies
- Top 10 Crypto Stocks - Blockchain leaders
- Top 10 Tech Stocks - Tech giants
