Google TPU v7 (Ironwood)
芯片
AI加速器
筹建中
Google's inference-age flagship: 4.6 PFLOPS FP8 and 9,216-chip superpods.
您正在查看一条芯片记录：该处理器是什么、其关键规格、谁设计它、谁制造它、谁供应其零部件，以及它被用在哪里。
Ironwood, announced April 2025 and generally available late 2025, is Google's first TPU pitched for the 'age of inference': 4,614 TFLOPS of FP8 per chip, 192 GB of HBM3E at 7.4 TB/s, and pods scaling to 9,216 liquid-cooled chips. Anthropic's October 2025 agreement to access up to one million TPUs — reported at up to a gigawatt of capacity in 2026 — made it the largest publicly known TPU commitment.
规格与关键参数
每个数据均标注来源依据：厂商规格、厂商声明、公开报道、估算或官方数据。"未公开披露"即字面含义——我们从不捏造数值。
Overview
Designer
Google
Announced
April 2025; GA late 2025
Status
Ramping
内存
Memory
192 GB HBM3EVendor spec
Memory bandwidth
7.37 TB/sVendor spec
算力
FP8
4,614 TFLOPSVendor spec
Physical
Pod scale
9,216 chips (superpod); ~42.5 FP8 ExaFLOPS per podVendor claim
Cooling
Direct liquid coolingVendor spec
它如何与整个技术栈相连
本记录的关联链接一览——点击分组可直接跳转。 +2 更多分组见下方。
内存技术
内部包含
支持方
来源
记录审阅截至 2026-01-15 — 查看 方法论.