# Alibaba has unveiled its latest AI chip, "Zhenwu M890" and AI LLM "Qwen3.7-Max" The chip is designed for Agentic AI workloads with a focus on AI inferencing.

*genai · news · 2026-05-20 · Wccftech*

## Key points

- Alibaba's Zhenwu M890 AI chip delivers 0.6 PFLOPs FP16 compute, outperforming NVIDIA Hopper H20 3x.
- M890 uses an in-house PPU architecture with a Transformer core engine and supports FP32, FP16, FP8, and FP4.
- The Panjiu AL128 Supernode Server integrates 128 AI accelerators per rack with PB/s scale bandwidth.
- Alibaba's Qwen3.7-Max model sustains 35-hour agentic operations and over 1,000 tool calls without degradation.
- Next year's Zhenwu V900 chip will offer 3x M890's performance, 216 GB memory, and 1200 GB/s bandwidth.

**Companies:** Alibaba
**Countries:** China

[Read the full story on Wccftech](https://wccftech.com/alibaba-targets-nvidia-hopper-with-zhenwu-m890-ai-chip-claiming-3x-h20-performance/)

---

Canonical: https://newsio.io/n/70b9bc6c-0494-4cc6-962d-98fa4bc9450d/alibaba-has-unveiled-its-latest-ai-chip-zhenwu-m890-and-ai-llm-qwen3-7-max-the-c
Summarized by Newsio from Wccftech. https://newsio.io/how-it-works