Alibaba unveils Zhenwu V900 AI chip and targets 20-gigawatt data center expansion
Alibaba has introduced the Zhenwu V900 AI accelerator, describing it as China's most powerful AI chip and claiming triple the performance of its predecessor. Alongside the chip launch, Alibaba announced targets to reach over 20 gigawatts of data center capacity by 2032 and plans to develop AI models scaling up to 10 trillion parameters.
Alibaba revealed the Zhenwu V900 chip at its annual Apsara Conference through its semiconductor division, T-Head. Pitched by the company as China's most powerful AI chip, the accelerator is designed for both model training and inference. Alibaba claims the processor delivers roughly three times the performance of its predecessor, the M890, which was introduced about four months earlier. The Zhenwu V900 features 216 gigabytes of memory, 1,200 gigabytes per second of bandwidth, and is designed to scale in clusters of up to 500,000 units ahead of expected mass production and commercial release in the first quarter of 2027.[1][3][7][10][11][12][13][16]
The chip launch accompanies a massive infrastructure roadmap in which Alibaba Cloud targets more than 20 gigawatts of global data center capacity by 2032. Backing this buildout, Alibaba has committed more than $53 billion over three years toward artificial intelligence infrastructure and development. In addition to the V900, the company plans to introduce Yitian 720 and 730 central processing units in 2027 tailored for agentic AI workloads.[1][4][6][8][9][14][15]
Alibaba also detailed ambitions for its Qwen large language models, stating that while Qwen 4 is currently in training, upcoming models including Qwen 4.5 and Qwen 5 are planned to reach between 5 trillion and 10 trillion parameters—a substantial jump from the 2.4 trillion parameters in Qwen 3.8 Max. The integrated push to develop proprietary chips and expand computing capacity is positioned as an effort to rival Nvidia and broaden domestic Chinese hardware alternatives. Following the announcements, Alibaba shares traded higher in premarket sessions.[1][2][4][8][9][11][17]
Key facts
- Alibaba's chip unit T-Head unveiled the Zhenwu V900 AI accelerator, claiming it is China's most powerful AI chip.
- Alibaba claims the Zhenwu V900 delivers roughly three times the performance of its M890 predecessor, which was released four months earlier.
- The Zhenwu V900 is planned for commercial release and mass production in Q1 2027 and is designed to support clusters of up to 500,000 units.
- Alibaba Cloud is targeting more than 20 gigawatts of global data center capacity by 2032.
- Alibaba has committed over $53 billion across three years toward AI capabilities and infrastructure.
- Alibaba plans to scale future Qwen AI models to between 5 trillion and 10 trillion parameters, compared to 2.4 trillion parameters in Qwen 3.8 Max.
- Alibaba plans to launch Yitian 720 and 730 CPUs for agentic AI workloads in 2027.
Sources · 14 sources
- WS
Wall St Engine@wallstenginePost on X ·
$BABA UNVEILS NEW AI CHIP, PLANS 5T-10T PARAMETER MODELS AND 20GW+ OF DATA CENTER CAPACITY Alibaba unveiled its new Zhenwu V900 AI chip, designed for both model training and inference, with performance roughly 3x higher than its predecessor released just four months ago. The V900 is expected to enter mass production and commercial release in Q1 2027. Alibaba says clusters built around the chip can scale to as many as 500,000 accelerators for training and running frontier AI models. Alibaba is also targeting more than 20GW of global data center capacity by 2032 and plans to launch its new Yitian 720 and 730 CPUs for agentic AI workloads in 2027. Qwen4 is currently in training and expected soon, while future Qwen generations are expected to scale to roughly 5T-10T parameters, versus 2.4T for the current Qwen 3.8 Max. Alibaba has committed more than $53B over three years toward AI infrastructure and capabilities.
Open source - AI
Analytics Insight@analyticsinmePost on X ·
𝐀𝐥𝐢𝐛𝐚𝐛𝐚 𝐒𝐭𝐨𝐜𝐤 𝐑𝐢𝐬𝐞𝐬 𝐚𝐬 𝐍𝐞𝐰 𝐀𝐈 𝐂𝐡𝐢𝐩 𝐁𝐨𝐨𝐬𝐭𝐬 𝐂𝐡𝐢𝐧𝐚 𝐓𝐞𝐜𝐡 𝐀𝐦𝐛𝐢𝐭𝐢𝐨𝐧𝐬 Alibaba is turning up the heat in the global AI race. The Chinese tech giant unveiled a new AI chip and flagship model, sending its shares higher in premarket trading. Its ambitious data-centre expansion plans also signal a push to strengthen AI infrastructure and compete with major US technology players. #Alibaba #AI #ArtificialIntelligence #Tech #StockMarket #China #AnalyticsInsight #AnalyticsInsightMagazine Read More 👇 https://t.co/d8DR2c2C9R
Open source - QU
Quartz@qzPost on X ·
Alibaba Just Unveiled China's Most Powerful AI Chip Alibaba dropped a bombshell at its annual Apsara Conference, revealing the Zhenwu V900 chip and plans for an AI model with up to 10 trillion parameters. The announcemen https://t.co/moo3lfMOr4 #Alibaba #ZhenwuV900 #AIChipRace https://t.co/CyWFJsVtRZ
Open source - FI
Finimize@finimizePost on X ·
What you need to know about markets today 👇 1️⃣ Alibaba and Tencent unveiled fresh AI products and plans, sending both stocks higher. Alibaba’s $53 billion spending push shows China isn’t easing off the AI accelerator – and could help narrow the gap if US rivals follow through on calls to slow development. 2️⃣ Polymarket wants European regulators to treat prediction markets as financial products rather than gambling, potentially opening a huge new market. But with Europe tightening betting rules – and Polymarket still fighting similar battles in the US – winning that argument could be tricky.
Open source - SC
South China Morning Post@SCMPNewsPost on X ·
Alibaba unveils ‘China’s top AI chip’, teases 10-trillion-parameter model https://t.co/vg1BspAM8Y
Open source - CO
Cointelegraph@CointelegraphPost on X ·
🇨🇳 NEW: China's Alibaba has unveiled its Zhenwu V900 AI chip to rival Nvidia, aiming to power 20GW of data center capacity by 2032. https://t.co/JfnLAqoCdJ
Open source - QU
Quartz@qzPost on X ·
Alibaba unveiled what it called China's most powerful AI chip and plans for a far larger model: The Zhenwu V900 delivers triple the performance of its predecessor and is set for commercial release in early 2027 https://t.co/ndLxqLiGVX https://t.co/NBRF5ie7Wv
Open source - WS
Wall St Engine@wallstenginePost on X ·
$BABA UNVEILS NEW AI CHIP, PLANS 5T-10T PARAMETER MODELS AND 20GW+ OF DATA CENTER CAPACITY Alibaba unveiled its new Zhenwu V900 AI chip, designed for both model training and inference, with performance roughly 3x higher than its predecessor released just four months ago. The V900 is expected to enter mass production and commercial release in Q1 2027. Alibaba says clusters built around the chip can scale to as many as 500,000 accelerators for training and running frontier AI models. Alibaba is also targeting more than 20GW of global data center capacity by 2032 and plans to launch its new Yitian 720 and 730 CPUs for agentic AI workloads in 2027. Qwen4 is currently in training and expected soon, while future Qwen generations are expected to scale to roughly 5T-10T parameters, versus 2.4T for the current Qwen 3.8 Max. Alibaba has committed more than $53B over three years toward AI infrastructure and capabilities as China pushes to reduce reliance on Nvidia and expand domestic chip production.
Open source - MT
MTS@MTSlivePost on X ·
SITUATION EXPLAINED: Alibaba published its AI roadmap. Qwen 5 will be 5 to 10 trillion parameters. • Qwen 4 is in training now, with Qwen 4.5 and Qwen 5 scaling to 5 to 10 trillion, roughly the size of Mythos, Fable, and Astra • Qwen 3.8 is around 2.4 trillion, so that's a 2 to 4x jump from where Chinese models sit today • Alibaba Cloud targets more than 20 gigawatts of data center capacity by 2032 • On recursive self-improvement, Qwen 3.8 Max completed 33 iterative cycles across a month of fully automated runs covering pipeline design, data validation, experimentation, and error diagnosis • A new training and inference chip claiming 3x the performance of the May version, plus next-gen CPUs built for agentic tasks • Chairman Joe Tsai: the total volume of machine thinking is less than 3% of all human thinking, with no explanation of how that was measured @theojaffee: "They're not at full recursive self-improvement yet. They haven't closed the loop, but they are getting much better at it."
Open source - TE
Techmeme@TechmemePost on X ·
Alibaba's T-Head unveils the Zhenwu V900 AI accelerator, which it says triples its predecessor's performance and can scale to clusters of up to 500,000 units (Bloomberg) (Visit Techmeme dot com for the link and full context!)
Open source - BL
Bloomberg@businessPost on X ·
Alibaba is rolling out what it calls China’s most powerful AI chip, an accelerator to compete with Nvidia and underpin a massive expansion of data center capacity in coming years https://t.co/Vq2lxOZxxH
Open source - 阳明
阳明AI@x_autonomyPost on X ·
【AI热点】14:00-16:00 更多详细信息→https://t.co/yRnv1SrjKv 1. 梁文锋称DeepSeek计划用华为芯片训练模型 [影响大·可执行低] 2. 阿里平头哥发布真武V900芯片，称算力达M890的3倍 [影响大·可执行中] 3. 美国加州州长签署7项法案严管AI数据中心 [影响大·可执行低] 4. 高瓴创投合伙人严文韬加入DeepSeek，终结三年CFO空缺 [影响中·可执行低] 5. 阿里千问平板QwenBook真机首秀：搭载小背屏，键盘设千问专属键 [影响中·可执行中] 6. 特斯拉FSD（监督版）获捷克批准，成为欧盟第七个开放国家 [影响中·可执行中] 7. 中国气象局发布妈祖气象智能预警方案中亚行动计划，尼泊尔定制版已交付 [影响小·可执行低] 8. D-RAC：通过PDF归一化与多模态Markdown转换实现企业文档通用检索感知摄取 [影响小·可执行高] 9. Tripo联手GPT-6 Astra打造UE5游戏场景 [影响中·可执行中] 10. oMLX作者Jun Kim加入Hugging Face，支持MLX社区 [影响小·可执行低] 11. X内置键盘快捷键总览，按？即可查看 [影响小·可执行高] 12. 特斯拉FSD v14.3.9出现软件漏洞，误报前置摄像头脏污 [影响小·可执行低] 13. CARE：面向视觉-语言-动作策略的经验引导原子纠错执行 [影响小·可执行中] 14. 1%的Token就够了：论在线策略蒸馏中的梯度估计 [影响小·可执行中] 15. 机械革命极光G505游戏主机上市：i7-14650HX+RTX 5060 Ti，首发9499元 [影响小·可执行高] 16. MiniCPM5-2B多智能体发票对账实测 [影响小·可执行高] 17. 小米18 Pro系列官宣首发传奇一瞬大模型影像，成像质感媲美专业相机 [影响中·可执行低] 18. 三星Galaxy A系列手机电池将植入安全识别芯片 [影响小·可执行低] 19. Jev因需求激增暂停新用户注册 [影响中·可执行低] 20. Gemini 4 Pro或即将发布 [影响大·可执行低] 21. Perplexity CEO：AI即操作系统，多数非AI应用将无头化 [影响中·可执行低] 22. Jev搭配Exa联网搜索效果惊人 [影响小·可执行高] 23. OpenAI预告今日将发布重大更新 [影响大·可执行中] 24. AI视频爱好者热情表白 [影响小·可执行低]
Open source - TH
The Hindu@the_hinduPost on X ·
The Alibaba Group unveiled its new Zhenwu V900 AI chip this week, pitching it as China’s most powerful, and promising tripled performance when compared to its predecessor. https://t.co/nbxVNlBVSj
Open source - BL
Bloomberg@businessPost on X ·
Alibaba unveils new AI chip it says is China's most powerful and targets 20 gigawatts of data center capacity by 2032. https://t.co/Gsh2HwT4sK 📷️: Qilai Shen/Bloomberg https://t.co/3mwwDmkGLm
Open source - CO
CoinGape@CoinGapeMediaPost on X ·
🇨🇳 NEW: @AlibabaGroup unveils its Zhenwu V900 AI chip, which it says delivers 3× the performance of its predecessor, as Alibaba Cloud targets more than 20GW of global data-centre capacity by 2032. https://t.co/0mOEy4SPv9
Open source - AI
Analytics Insight@analyticsinmePost on X ·
𝗔𝗹𝗶𝗯𝗮𝗯𝗮 𝗨𝗻𝘃𝗲𝗶𝗹𝘀 𝗭𝗵𝗲𝗻𝘄𝘂 𝗩𝟵𝟬𝟬 𝗔𝗜 𝗖𝗵𝗶𝗽 𝘄𝗶𝘁𝗵 𝟯𝗫 𝗣𝗲𝗿𝗳𝗼𝗿𝗺𝗮𝗻𝗰𝗲 𝗖𝗹𝗮𝗶𝗺: 𝗪𝗵𝘆 𝗶𝘁 𝗠𝗮𝘁𝘁𝗲𝗿𝘀 Alibaba unveiled its Zhenwu V900 AI chip, claiming 3X the performance of its M890 predecessor. With 216GB memory, 1,200GB/s bandwidth and support for clusters of up to 500,000 chips, the V900 is set for commercial release in Q1 2027. #Alibaba #ZhenwuV900 #AIChip #ArtificialIntelligence #AIInfrastructure #THead #Qwen #AnalyticsInsightUAE Read More 👇 https://t.co/akF6eaShTX
Open source - MI
Mint@livemintPost on X ·
Alibaba unveils new AI chip to challenge NVIDIA, plans Qwen models with up to 10 trillion parameters https://t.co/kC5J1iCm8M
Open source - SB
Shay Boloor@StockSavvyShayPost on X ·
$BABA just unveiled a new AI chip while laying out plans for 5T to 10T parameter models and more than 20GW of data center capacity by 2032. That roadmap is now backed by more than $53B of AI investment over three years. https://t.co/a192D6H0bn
Open source