Alibaba Group has unveiled a new processor designed for autonomous digital agents, marking the Chinese tech giant’s latest push to reduce dependence on U.S. chipmakers amid tightening export restrictions.
The company introduced the Zhenwu M890 on Wednesday at its annual cloud summit in China. Developed by Alibaba’s chip unit T-Head, the processor delivers roughly three times the performance of the company’s previous-generation Zhenwu 810E chip.
Alibaba said the chip was built specifically for workloads tied to autonomous software agents that can complete complex tasks with minimal human input. These systems typically require long memory retention and constant communication between models operating simultaneously.
The launch comes as Chinese technology firms accelerate efforts to develop domestic alternatives to processors made by Nvidia following U.S. restrictions on advanced semiconductor exports to China.
Unlike traditional processors optimized mainly for training large language models, the M890 is aimed at handling continuous agent operations where systems must coordinate actions, retain context for extended periods, and process large volumes of data in real time.
Alibaba also revealed a long-term semiconductor roadmap extending through 2028. The company said the next-generation V900 chip, expected in 2027, would deliver another threefold performance increase over the M890. A third processor, called the J900, is planned for 2028.
The roadmap signals Alibaba’s intention to build a sustained in-house chip ecosystem as competition intensifies in China’s computing sector.
The announcement also reflects a broader shift in the computing industry toward agent-based systems that can independently perform extended tasks instead of responding to single prompts. Such workloads demand far greater memory capacity, faster communication between processors, and continuous uptime, pushing companies to redesign hardware specifically for long-running autonomous operations.
Last year, Alibaba pledged to invest more than 380 billion yuan, or about $53 billion, into cloud and computing infrastructure over a three-year period. The company said demand for agent-based systems is expected to increase sharply as businesses adopt more autonomous software tools.
Alongside the new chip, Alibaba unveiled the Panjiu AL128 server system, which combines 128 accelerators into a single rack for enterprise-scale deployments.
The company said the server platform is already available to domestic customers through its Bailian cloud platform.
T-Head also revealed that more than 560,000 Zhenwu processors have been shipped so far, with over 400 external customers across 20 industries deploying the chips. Those sectors include financial services and automotive manufacturing.
Alibaba additionally launched Qwen 3.7-Max, the latest version of its flagship large language model. The company said the model was designed for advanced coding and long-duration agent tasks and can operate continuously for up to 35 hours without performance degradation.
The announcement highlights how China’s largest technology companies are increasingly investing in their own processors as Washington tightens controls on access to advanced American semiconductor technology.
Chinese firms, including Huawei, have also accelerated the development of domestic computing hardware in response to export restrictions.
Unlike earlier waves of computing infrastructure focused primarily on chatbot systems, Alibaba’s latest hardware strategy centers on autonomous software agents capable of carrying out extended, multi-step operations with limited supervision.
Get the latest in engineering, tech, space & science - delivered daily to your inbox.
With over a decade-long career in journalism, Neetika Walter has worked with The Economic Times, ANI, and Hindustan Times, covering politics, business, technology, and the clean energy sector. Passionate about contemporary culture, books, poetry, and storytelling, she brings depth and insight to her writing. When she isn’t chasing stories, she’s likely lost in a book or enjoying the company of her dogs.
此内容由惯性聚合(RSS阅读器)自动聚合整理，仅供阅读参考。 原文来自 — 版权归原作者所有。