한국어EN
Inference Infrastructure & ServingGoogle Cloud

Google Cloud announces Ironwood (TPU v7)

Made by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made

Announced Ironwood (TPU v7) — 9,216 chips per superpod

Why it matters

A TPU for the inference era. Actual GA came on Mar 31, 2026

Source — Primary sources first; secondary reports are labelled “Press”

Officialcloud.google.comhttps://cloud.google.com/blog/products/compute/ironwood-tpus-and-new-axion-based-vms-for-your-ai-workloads/

Added in Week 35, 2026 · See that week

On the map — ⑧ Inference Infrastructure line

  1. Previous stopDeepSeek-V3.2-Exp
  2. This record · stopIronwood
  3. Next stopGoogle Cloud previews its 8th-generation TPUs at Cloud Next
See it on the map

Same week

Google Cloud adds Auto Vector Embeddings and Auto Vector Index to AlloyDB AIOfficialTerminal-Bench team releases Terminal-Bench 2.0 and Harbor evaluation harnessOfficialMoonshot AI releases Kimi K2 ThinkingOfficialAnthropic publishes Code Execution with MCPOfficial

Nearby on the same axis

NVIDIA announces DGX Spark (ships Oct 15)OfficialDeepSeek-V3.2-ExpOfficialNVIDIA takes non-exclusive Groq inference license, hires Jonathan Ross and teamOfficialFlashAttention-4Paper