TREE NEWS update: Alibaba Cloud unveiled its next-generation Cloud Parallel File Storage (CPFS) at the 2026 Yunqi Conference, offering throughput in the hundreds of TB/s and billions of IOPS, with single-file-system scale raised fivefold to 100 PiB. In training, the system cuts average model startup time by 50%, lifts peak compute utilization by 30% and reduces AI storage costs by 69%, doubling workload capacity.
Alibaba Cloud launches CPFS storage, cutting AI storage costs 69%
AI take
The headline number is cost, but the more consequential claim is utilization: squeezing 30% more out of existing compute reframes storage as a lever on GPU economics, not just a line item. That matters most for large model trainers, where startup time and idle accelerators are the real budget drain. The open question is whether these gains hold beyond Alibaba's own stack, or remain a bundled advantage for customers already inside its cloud.
Generated by AI for reference only.
Share on WeChat
Open WeChat → Scan → then tap "…" to send to a chat or Moments.
Tap "…" in the top-right corner to send to a chat or share to Moments.
Related News
13m ago
ByteDance’s Doubao Cuts General Session Team by About Half
31m ago
Shanghai Regulators Hold First Multi-Party Talks on Online Info Governance for Pioneer Industries
38m ago
UN chief Guterres warns AI risks power shifting to private firms
42m ago
Alibaba’s Qwen Office Launches QwenNote AI Hardware, A2 Priced at 1,199 Yuan
1h ago
XPeng Group Holds First Robot Supply Chain Partner Conference
1h ago
Hillhouse Venture Capital ex-partner joins DeepSeek as CFO