Pyyan / Compare / MI400 vs B200 vs H100 vs TPU v7 Ironwood

MI400 vs B200 vs H100 vs TPU v7 Ironwood

4 of 5

GPU & AI Chips · verified 13 Aug 2026

×MI400AMDcurrent
×B200NVIDIAcurrent
×H100NVIDIAcurrent
×TPU v7 IronwoodGooglecurrent
1 slot left
SpecificationMI400B200H100TPU v7 Ironwood
SummaryDouble NVIDIA's memory capacity.The current volume Blackwell part.The workhorse of the previous generation.Inference-oriented, Google Cloud only.
Memory432GB HBM4192GB HBM3e80GB HBM3192GB HBM3e
Bandwidth19.6 TB/s7.37 TB/s
Price$30–50K$25–40K
VendorAMDNVIDIANVIDIAGoogle
CategoryGPU & AI ChipsGPU & AI ChipsGPU & AI ChipsGPU & AI Chips
OfficialAMDNVIDIANVIDIAGoogle

Highlighted rows are where these differ.

MI400

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

Best for large-model inference.

Full spec sheet →

B200

  • Sold out through mid-2026
  • ~3.6M unit backlog

Best for general training.

Full spec sheet →

H100

The workhorse of the previous generation.

Best for value.

Full spec sheet →

TPU v7 Ironwood

  • Requires JAX
  • Not purchasable or on any marketplace

Best for GCP inference.

Full spec sheet →