PyyanThe AI Index

Pyyan / Compare / H100 vs MI400 vs TPU v7 Ironwood

H100 vs MI400 vs TPU v7 Ironwood

3 of 5

GPU & AI Chips · verified 13 Aug 2026

×H100NVIDIAcurrent
×MI400AMDcurrent
×TPU v7 IronwoodGooglecurrent
2 slots left
SpecificationH100MI400TPU v7 Ironwood
SummaryThe workhorse of the previous generation.Double NVIDIA's memory capacity.Inference-oriented, Google Cloud only.
Memory80GB HBM3432GB HBM4192GB HBM3e
Bandwidth—19.6 TB/s7.37 TB/s
Price$25–40K——
VendorNVIDIAAMDGoogle
CategoryGPU & AI ChipsGPU & AI ChipsGPU & AI Chips
OfficialNVIDIA ↗AMD ↗Google ↗

Highlighted rows are where these differ.

H100

The workhorse of the previous generation.

Best for value.

Full spec sheet →

MI400

  • 19.6 TB/s bandwidth
  • 40 PFLOPS FP4
  • $7.2B in commitments

Best for large-model inference.

Full spec sheet →

TPU v7 Ironwood

  • Requires JAX
  • Not purchasable or on any marketplace

Best for GCP inference.

Full spec sheet →