← All hardware
Google TPU v7 (Ironwood) logo
hardware Data Center AI Accelerators

Google TPU v7 (Ironwood)

by Google

Google's 7th-gen TPU built from the ground up for the age of inference.

Pros

  • Massive 9,216-chip single-pod scale
  • Excellent perf/watt and price/performance via cloud
  • Deep integration with Google Cloud and JAX

Cons

  • Cloud-only; cannot be purchased on-prem
  • Locked to Google Cloud ecosystem
  • Smaller third-party software ecosystem than CUDA

✓ Where it shines / best for

  • Large-scale, latency-sensitive LLM inference serving
  • Frontier-model and Mixture-of-Experts deployment on Google Cloud
  • Enterprises needing high memory-capacity-per-chip for big models

✕ Not the best fit for

  • On-premises hardware buyers (cloud-only)
  • Tiny or hobbyist workloads
  • Teams requiring native CUDA ecosystem

Features

  • ✓ LLM
  • ✓ API access
  • ✓ HBM3E
  • ✓ Inference
  • ✓ Training
  • ✓ High Bandwidth
  • ✓ Jax
  • ✓ Cloud TPU
  • ✓ Moe

Pricing

PlanPriceBillingNotes
Cloud rental (GA pricing)Contact Google Cloudper chip-hourIronwood (7th-gen TPU) offered via Google Cloud; public per-chip-hour list pricing limited at launch — quoted through sales/committed-use.
Committed-use discountsvaries1-yr / 3-yrStandard Google Cloud CUD discounts apply once generally available.

Pricing verified from the official source. Prices change often — confirm on the vendor's site before buying.

Specifications

memory192 GB HBM per chip, ~7.37 TB/s
coolingliquid-cooled, cloud-only
pod_scale9,216 chips per pod
architectureIronwood (TPU v7 / tpu7x), Google custom ASIC
perf_per_watt2x vs TPU v6e Trillium
fp8_performance~4.6 PFLOPS peak per chip
Sponsored

A full review is being generated for this product and will appear here shortly.

Compare with

Compare
Compare