- New chip set to drive AI, cloud computing, and robotics markets with 3x better performance than predecessor
- Launch signals challenge to global CPU giants with breakthrough open-source architecture
Alibaba’s DAMO Academy, its research and development arm, introduced its next-generation flagship CPU, the XuanTie C950, at the 2026 XuanTie RISC-V Ecosystem Conference in Shanghai on March 24, marking a significant milestone in the rise of RISC-V architecture.
The chip achieved a breakthrough single-core performance score of over 70 in the SPECint2006 benchmark, surpassing records of previous RISC-V processor series.
Designed for cloud computing, generative AI, high-end robotics, and edge computing, the C950 offers more than three times the performance of its predecessor, the C920.
The chip’s architecture leverages the open-source, flexible nature of RISC-V, allowing it to be customized for a range of applications.
Notably, the C950 integrates a self-developed AI acceleration engine and natively supports large-scale models such as Qwen3 and DeepSeek V3, positioning it as a key player in the emerging AI Agent era.
RISC-V’s open-source framework contrasts with traditional closed-source models like x86 and ARM, and it is increasingly seen as a future-proof solution for AI-driven hardware.
According to Ni Guangnan, an academician of the Chinese Academy of Engineering, RISC-V now holds a 25% share of the global processor market, signaling the loosening grip of the x86 and ARM duopoly.
The XuanTie C950’s ability to run AI models with hundreds of billions of parameters without licensing fees further strengthens the cost-efficiency argument for RISC-V in the AI inference market.
This aligns with predictions of rising demand for affordable AI processing power. The chip is expected to integrate smoothly into Alibaba’s broader ecosystem, providing a critical foundation for its cloud and AI services.
Public records show that the DAMO Academy has become a global leader in RISC-V development, having released more than 10 XuanTie CPU models and deploying them in thousands of terminal products across industries like servers, robotics, new energy vehicles, AI endpoints, and storage controllers.
