💰钛媒体•Mar 24, 2026Stalecollected in 11m

Alibaba DAMO CPU Natively Supports Qwen3

Post LinkedIn

💰Read original on 钛媒体

#ai-cpu #llm-hardware #chip-launchdamo-cpu

💡Alibaba's first Qwen3-native CPU boosts domestic LLM hardware sovereignty

⚡ 30-Second TL;DR

What Changed

DAMO Academy launched new CPU

Why It Matters

Bolsters Alibaba's AI stack with custom silicon, reducing reliance on foreign chips and accelerating Qwen deployments in China. Positions Alibaba competitively in AI infrastructure race.

What To Do Next

Benchmark Qwen3 inference on Alibaba Cloud's new DAMO CPU via their console.

Who should care:Enterprise & Security Teams

🧠 Deep Insight

Web-grounded analysis with 9 cited sources.

🔑 Enhanced Key Takeaways

•The new CPU, named Xuantie C950, is built on the open-source RISC-V architecture and utilizes a 5nm process, achieving a clock speed of up to 3.2 GHz.
•The Xuantie C950 achieved a single-core general-purpose performance score exceeding 70 points in the SPECint2006 benchmark, which Alibaba claims is a new global performance record for RISC-V CPUs and over three times the performance of its predecessor, the C920.
•Beyond Qwen3, the processor's proprietary AI acceleration engine also provides native support for other large-scale models with hundreds of billions of parameters, such as DeepSeek V3, targeting applications in cloud computing, high-end robotics, and edge AI agents.

🛠️ Technical Deep Dive

•Architecture: RISC-V (Open-source)
•Process Node: 5nm
•Clock Speed: Up to 3.2 GHz
•Microarchitecture: 8-instruction decoding, 16-stage pipeline, and an out-of-order window of over 1,000 instructions
•Benchmark: >70 points in SPECint2006 (single-core)
•AI Acceleration: Integrated proprietary AI acceleration engine for native LLM inference

🔮 Future ImplicationsAI analysis grounded in cited sources

RISC-V will capture a larger share of the high-performance AI computing market.

The Xuantie C950's ability to run massive LLMs natively demonstrates that RISC-V can compete with established x86 and Arm architectures in high-end AI workloads.

Alibaba will accelerate the integration of AI models directly into hardware ecosystems.

The strategic focus on 'chip-model synergy' and the expansion of the 'Wujian Alliance' suggest a shift toward hardware-software co-design for AI agents.