Alibaba's chipmaking unit T-Head unveils the Zhenwu M890 for AI training and inference, saying it is well-suited for agentic tasks, and plans annual upgrades
Bloomberg Luz Ding
Related Coverage
- Alibaba Takes Aim At Nvidia With New 3X Powerful AI Chip And Next-Gen LLM Model As China Tech War Heats Up Benzinga · Namrata Sen
- Alibaba unveils new AI chip in push for domestic alternatives Reuters
- Alibaba Unveils New AI Chip as China Accelerates Adoption of Domestic Chips The Information · Juro Osawa
- Is Alibaba's new AI chip timed to steal Nvidia's thunder? Proactive · Ian Lyall
- Alibaba Unveils New AI Chip, Upgrades AI Model Wall Street Journal · Tracy Qu
- Alibaba Ramps Up AI Push With New Chip, Model Upgrades — Update Dow Jones Newswires · Tracy Qu
- China's Alibaba unveils new AI chip, LLM in push for domestic alternatives Business Standard
- Alibaba's T-Head doubles down on AI infrastructure with Zhenwu M890 DigiTimes
- Alibaba unveils the Zhenwu M890 as China's NVIDIA alternative push hardens The Next Web · Allison Steffens Herrera
- Alibaba (BABA) Launches Zhenwu M890 Chip to Challenge Nvidia's AI Dominance Blockonomi · Trader Edge
- Alibaba Unveils New AI Chip, Flagship Model, and Rebuilt Cloud Stack AI for Agentic Era Alizila · Crystal Liu
- Qwen3.7: The Agent Frontier Qwen
- Alibaba unveils new AI chip in push for domestic alternatives Reuters
- Alibaba is designing AI chips around agents, and that changes what the race is actually about AI News · Dashveenjit Kaur
- Qwen3.7-Max: The Agent Frontier Hacker News
- Alibaba releases AI chip for training, inference Tech in Asia · Grace Priscilla Teo
- Alibaba Launches Zhenwu M890 AI Chip in China Push WinBuzzer · Markus Kasanmascheff
- Alibaba Unveils AI Chip in Bid to Counter NVIDIA and U.S. Export Curbs Techstrong.ai · Jon Swartz
- AMD: An agentic AI and inference winner The Motley Fool · Geoffrey Seiler
Discussion
-
@poezhao0605
Poe Zhao
on x
T-Head revealed a four-generation Zhenwu chip roadmap through 2028. 810E (2024) → M890 (2026 Q2) → V900 (2027 Q3) → J900 (2028 Q3). Memory scales from 96GB to 216GB. Each generation claims 3x performance gains on a proprietary parallel computing architecture. This is the [image]
-
@hexxrl
@hexxrl
on x
new models and training/inference chips from $BABA Zhenwu M890 delivers 3x the performance vs Zhenwu 810E same event they revealed qwen3.7 max which can run continuously for 35 hours without degradation if baba was a western company it would be worth $2T plus [image]
-
@poezhao0605
Poe Zhao
on x
Alibaba Cloud announced two new T-Head chips at its summit: Zhenwu M890, a unified training and inference AI processor, and ICNSwitch, a dedicated interconnect chip. A standalone switch chip signals cluster-scale ambition. [image]
-
@stockmktnewz
Evan
on x
Alibaba $BABA just announced today a new AI chips that is says is 3x more powerful than its last chip “The Zhenwu M890 delivers three times the performance of the current Zhenwu 810E, Alibaba said, adding that the new processor has 144 GB GPU memory and interchip bandwidth of [im…
-
@tphuang
@tphuang
on x
Alibaba unveiled Zhenwu M890 AI chip today along w/ 128-chip .久AL128 SuperNode, ICN Switch 1.0 as part of its AI hardware upgrade. Will now use 800G interconnect & 144GB of memory/chip. Supports FP4 to FP32 (in line w/ Ascend) Expect to release V900 & J900 chip in next 2 yrs [ima…