技术档案
液冷AI服务器Liquid-cooled AI servers
用液体散热的AI服务器
每条都摘自下方编号的来源原文；只写事实，不写预测或建议。
技术简介
- 液冷通过冷板、歧管至冷却塔的完整液体分配回路移走热量。1
With far greater AI operations per second comes a greater heat load: heat must be moved through the full fluid distribution loop, from the cold plates through manifolds, to the cooling tower.
已有应用
- Vera Rubin NVL72机架把72颗NVIDIA Rubin GPU和36颗NVIDIA Vera CPU置于一个液冷机架中。1
72 NVIDIA Rubin GPUs and 36 NVIDIA Vera CPUs in a single liquid-cooled rack operating as one machine; Eighteen 1U compute trays, each with four Rubin GPUs and two Vera CPUs, connect through nine sixth-generation NVIDIA NVLink switch trays delivering 216 TB/s of scale-up bandwidth; Each rack provides 20.7 TB of HBM4 memory and up to 54 TB of LPDDR5X
主要参与者
- Supermicro提供冷板、歧管、CDU和冷却塔等液冷组件。1
Supermicro's DCBBS provides unmatched thermal expertise and a comprehensive in-house portfolio of liquid cooling components.
- Supermicro的液冷组件组合包含冷板、歧管、CDU、后门换热器和冷却塔。1
Cold plates, manifolds, hose kits, rack power shelves, in-row cooling distribution units (CDUs), in-rack CDUs, Liquid-to-Air sidecar CDUs, rear door heat exchangers, and facility-side cooling towers all come from the same portfolio.
相关卡片
- 2026-10-20客户买服务器等产品，服务器厂商超微季度营收 199 亿美元，去年同季 50 亿