Alibaba Unveils Zhenwu V900 Chip and 5-10 Trillion Parameter Model Plan at Apsara Conference

At the 2026 Apsara Conference, Alibaba CEO Eddie Wu announced a full-stack AI strategy, while Reuters reports the company is planning to train massive models with 5 to 10 trillion parameters. Doubling down on both chips and models, the move is seen as an aggressive bet by a Chinese vendor on the frontier AI race.

Confirmed

• Alibaba released its in-house AI chip Zhenwu V900, billed by the company as China's most powerful AI chip, though detailed specs and performance figures have not been disclosed
• A single cluster can scale to 500,000 cards working in concert
• According to Reuters, Alibaba plans to train models with 5 to 10 trillion parameters; leaks suggest the massive scale is intended for Qwen 4.5 and Qwen 5, and Qwen 4 is already in training
• Multiple posts mention artificial superintelligence (ASI) as Alibaba's long-term goal, with a model roadmap pushing toward recursive self-improvement (RSI)

Unconfirmed

• Zhenwu V900's specific performance specs, interconnect solution, and mass-production timeline remain undisclosed
• The training timeline, compute sources, and final form of the 5-10 trillion parameter models are still at the planning/rumor stage; Alibaba has not provided full details

Why it matters

If realized, 5-10 trillion parameters would far exceed today's mainstream frontier models, showing Alibaba pushing hard on both compute (in-house chips) and models (ultra-large-scale training). The 500,000-card single-cluster scalability also reflects a new approach by Chinese vendors to the compute bottleneck of large-scale training—worth watching as official details and actual training progress emerge.

2026-09-22 ~ 2026-09-22 · 17 related posts

• Alibaba Plans AI Model With 5-10 Trillion Parameters, Unveils New Chip — tengo_harambe · 2026-09-22
• Alibaba roadmap: Qwen explores RSI, new models up to 10T params, 20GW datacenters by 2032 — teortaxesTex · 2026-09-22
• Alibaba unveils new AI chip it calls China's most powerful — AIFlow_ML · 2026-09-22
• [source] Alibaba reported targeting 5-10T-parameter Qwen models, training Qwen 4 with 20 GW datacenter plan — mark_k · 2026-09-22
• Alibaba Unveils Zhenwu V900, China's Most Powerful AI Chip, Scaling to 500K Cards — CurieuxExplorer · 2026-09-22
• Alibaba bets full-stack on AI: 5-10T param Qwen, Zhenwu V900 chip, 20GW cloud by 2032 — Scobleizer · 2026-09-22
• [source] Alibaba CEO Eddie Wu says new AI model to train on 5-10 trillion parameters — JosephJacks_ · 2026-09-22
• Alibaba lays out AI roadmap: 10T-parameter Qwen, Zhenwu V900 chip, 20GW data centers by 2032 — iamfakhrealam · 2026-09-22
• Alibaba CEO Eddie Wu: Machine thinking is under 3% of human thinking, with 1000x runway ahead — alifcoder · 2026-09-22
• Qwen rumored to push 5-10T params with recursive self-improvement, plus Yunqi lineup — lxfater · 2026-09-22
• Qwen4 training now: Alibaba teases Max/Flash/Plus variants, 5-10T params for Qwen4.5-5 — ChrisGPT · 2026-09-22
• [source] Alibaba Cloud Summit: Zhenwu V900 chip, Qwen4 teaser and a dozen AI launches — 智东西 · 2026-09-22
• Alibaba unveils new powerful chip and ambitious AI model plans — boppinmule · 2026-09-22
• Alibaba Roadmap: 20GW Global Compute by 2032, Zhenwu Chips 3x Faster — kevinsxu · 2026-09-22
• Alibaba's Qwen 3.8 Max Runs on All-Domestic Zhenwu M890 Chips, Reports Claim — bookwormengr · 2026-09-22

2 near-duplicate retellings: iamfakhrealam · robleclerc