惯性聚合 高效追踪和阅读你感兴趣的博客、新闻、科技资讯
阅读原文 在惯性聚合中打开

推荐订阅源

C
Cisco Blogs
爱范儿
爱范儿
有赞技术团队
有赞技术团队
博客园 - 【当耐特】
Jina AI
Jina AI
Project Zero
Project Zero
宝玉的分享
宝玉的分享
Martin Fowler
Martin Fowler
WordPress大学
WordPress大学
Simon Willison's Weblog
Simon Willison's Weblog
钛媒体:引领未来商业与生活新知
钛媒体:引领未来商业与生活新知
T
Tenable Blog
F
Fortinet All Blogs
大猫的无限游戏
大猫的无限游戏
Last Week in AI
Last Week in AI
月光博客
月光博客
雷峰网
雷峰网
G
Google Developers Blog
V
V2EX
T
Tor Project blog
罗磊的独立博客
Schneier on Security
Schneier on Security
Know Your Adversary
Know Your Adversary
W
WeLiveSecurity
freeCodeCamp Programming Tutorials: Python, JavaScript, Git & More
P
Privacy International News Feed
S
Securelist
奇客Solidot–传递最新科技情报
奇客Solidot–传递最新科技情报
P
Proofpoint News Feed
Blog — PlanetScale
Blog — PlanetScale
Threat Intelligence Blog | Flashpoint
Threat Intelligence Blog | Flashpoint
小众软件
小众软件
Scott Helme
Scott Helme
I
Intezer
T
Threat Research - Cisco Blogs
The GitHub Blog
The GitHub Blog
N
Netflix TechBlog - Medium
C
CERT Recently Published Vulnerability Notes
Security Archives - TechRepublic
Security Archives - TechRepublic
酷 壳 – CoolShell
酷 壳 – CoolShell
L
LINUX DO - 最新话题
N
News | PayPal Newsroom
L
Lohrmann on Cybersecurity
T
Troy Hunt's Blog
Google DeepMind News
Google DeepMind News
P
Proofpoint News Feed
人人都是产品经理
人人都是产品经理
Latest news
Latest news
AWS News Blog
AWS News Blog
Apple Machine Learning Research
Apple Machine Learning Research

The Register - On-Prem: Systems

Qualcomm teases agentic CPUs and smartphones Fujitsu says quantum and AI will replace mainframes in 2035 ZTE & China's NCRC partner for smart interventional medicine Core Scientific accelerates crypto-to-AI pivot Meta to use millions of AWS Graviton cores AI now gobbling up power and management chips for servers Tesla stakes AI dreams on Intel's unfinished AI chip SK Hynix breaks ground on Indiana advanced packaging plant Datacenter boom keeps dirty coal plants alive in the US AMD's Ryzen 9 9950X3D2 Dual Edition tested World's blandest man steps down from CEO job to spend more time in tastefully appointed home World's blandest man steps down from CEO job Intel eases reliance on TSMC with 'Merica-made Core Series 3 processors Intel eases reliance on TSMC with Core Series 3 CPUs Guide to GPU virtualization: passthrough, vGPU, and MIG Guide to GPU virtualization: passthrough, vGPU, and MIG Orbital datacenter startup admits launch economics don't fly AI-powered mainframe exits are a bubble set to pop Cloud-smart strategy helps Interactive meet GenAI demands Oracle taps Bloom for fuel cells to support datacenter binge Oracle taps Bloom for fuel cells to support datacenter binge UK signs Rolls-Royce SMR design deal Japan going back to the future by reviving its chip industry Japan going back to the future by reviving its chip industry AWS ponders selling its home-grown chips by the rack-load Supply chain challenges risk delaying Nvidia's Rubin GPUs Supply chain challenges risk delaying Nvidia's Rubin GPUs Supermicro investigating alleged China chip smuggling Intel trapped in Elon's reality distortion field UALink delivers 2.0 spec before v. 1.0 silicon ships OpenInfra General Manager on sovereignty and kill switches Anthropic reveals $30bn run rate, plan to use new Google TPU Nvidia embraces optical scale-up as copper reaches limits Nvidia embraces optical scale-up as copper reaches limits IBM wants Arm software on its mainframes for AI support AI datacenters create heat islands around them, paper finds Arm says AI agents need a new CPU. Intel doesn't buy it Memory-makers' shares are down. Don't blame Google Memory-makers' shares are down. Don't blame Google US PC shipments to fall 13% as memory and storage crunch hits budget systems US PC shipments to fall 13% as memory costs surge ZTE showcases intelligent computing at CloudFest 2026 Rebellions eyes global expansion with rack-scale AI platform Rebellions eyes global expansion with rack-scale AI platform Enterprise infrastructure is entering an economic reset AMD doubles up on V-Cache with 9950X3D2 Dual Edition Apple's making more iPhone bits in US, but not the iPhone Three more charged with trying to smuggle GPUs to China Three more charged with trying to smuggle GPUs to China Dell slims Pro laptops, boosts battery and cooling Alibaba delivers RISC-V server chip optimized for Chinese AI AI-pilled Arm CEO teases mystery products for $1T TAM Arm rolls its own 136-core AGI CPU to chase AI hype train SoftBank builds AI mega-datacenter on nuke site SoftBank builds AI mega-datacenter on nuke site Chip tester shrugged off ransomware – then came the leak Explainer: AI-ready servers Elon Musk proposes 'Terafab' to level up chip production Australia to datacenter operators: BYO energy or stay home Australia to datacenter operators: BYO energy or stay home Supermicro co-founder charged over $2.5B GPU sales to China Blue Origin applies to launch 51,000 datacenter satellites Blue Origin applies to launch 51,000 datacenter satellites Alibaba has made 470,000 AI chips, admits they’re inferior Decoding Nvidia's Groq-powered LPX and the rest of its new rack systems Your next car might need 300 GB of RAM, and so will autonomous robots Your next car might need 300 GB of RAM, and so will robots It's not a binary choice: Boffin builds ternary CPU Nvidia H200 back on in China, production ramping: Huang Nvidia slaps $20B Groq tech into massive new LPX racks to speed AI response time AI Burning Man happens next week – what to expect at Nvidia GTC 2026 Meta reveals four Broadcom-built custom AI chips, claims some outperform commercial silicon Meta reveals custom AI chips it says beat Nvidia ZTE and Orange Morocco launch Livebox 7 for smart homes Ayar Labs taps Wiwynn to cram 1,024 GPUs into a photonic rack system Ayar Labs, Wiwynn to cram 1,024 GPUs into photonic system ZTE and Whale Cloud Showcase Digital Transformation at MWC AI datacenters may gulp NYC's daily water supply at peak Mystery outage behind JetBlue's request for grounding HPE tweaks T&Cs so it can change quotes as RAM prices rise Supermicro launches probe after staff charged with China export violations
Alibaba delivers RISC-V server chip optimized for Chinese AI
Simon Sharwood Simon Sharwood · 2026-03-25 · via The Register - On-Prem: Systems

Systems

Alibaba delivers RISC-V server chip optimized to run China’s top AI models

Claims its set performance records but looks to be years behind western fare

Alibaba has revealed a new server chip that it says is the most powerful processor ever to use the RISC-V instruction set.

According to a social media post by Alibaba’s DAMO Academy, which develops some of its chips, the new XuanTie C950 is ready to power cloudy servers, generative AI workloads, high-end robotics, and edge computing devices.

“The XuanTie C950 is equipped with a self-developed AI acceleration engine, and for the first time natively supports large models with hundreds of billions of parameters, such as Qwen3 and DeepSeek V3, potentially becoming a new type of high-end CPU for the AI Agent era,” the post enthuses.

Alibaba claims the machine’s single-core general-purpose performance “exceeded 70 points in the SPECint 2006 benchmark test.” Photos from Tuesday event at which Alibaba announced the chip suggest its SPECInt 2017 benchmark result is 2.6GHz. Per analysis by Google researcher Laurie Kirk puts it nearly on par with Apple’s M1 chip – which the iGiant launched in the year 2020.

Kirk also noted that Alibaba claimed it’s implemented version 23.1 of the RISC-V RVA, a minor update proposed in August 2025. She expressed surprise Alibaba was able to use it so quickly.

A product page and spec sheet for the chip add more details – but also omit details like core count and instead offer the non-specific description of the chip as a “64-bit multi-core CPU IP.”

Based on those documents the “acceleration engine” mentioned above is probably the XuanTie Tensor Processing Engine (TPE) and supports data types from FP16 down to INT4/FP8, plus the micro-scaling formats MXFP8, MXFP4, and RVFP4. The chip can achieve 8 TOPS per TPE.

The memory subsystem apparently “comprises a high-performance multi-level cache hierarchy with an ultralow 4-cycle load-to-use L1 data cache latency, a private per-core L2 cache supporting large capacity configurations, and an MMU with multiple RISC-V virtual memory modes and two-stage address translation.”

In a section of the spec sheet discussing buses, there’s a mention of a multi-processor mode that “leverages the XL-300 interconnect to form clusters of up to 8 cores.”

Posts The Register has seen suggest Alibaba had the chip made using a 5nm process, a feat that’s not beyond some Chinese chipmakers. Whether they can build chips of that caliber in large quantities is less certain.

Alibaba CEO Yongming Wu last week acknowledged that Chinese chips lag behind those from western chipmakers. He said the company’s response is “to engage in more profound co-design with Alibaba's cloud infrastructure and the Qwen model to provide improved cost effectiveness. This is one key differentiator … that sets us apart from other chip companies.”

So while the XuanTie C950 may not be a particularly powerful chip, and Alibaba may struggle to get millions of them into production, native support for the company’s own Qwen models suggests Wu’s vision for a nicely harmonized AI stack is becoming a reality. ®