PyyanThe AI Index

Pyyan / Compare / Trainium 3 vs B300 Blackwell Ultra vs H100

Trainium 3 vs B300 Blackwell Ultra vs H100

3 of 5

GPU & AI Chips · verified 13 Aug 2026

×Trainium 3AWScurrent
×B300 Blackwell UltraNVIDIAcurrent
×H100NVIDIAcurrent
2 slots left
SpecificationTrainium 3B300 Blackwell UltraH100
Summary30–40% better price-performance.15 PFLOPS dense FP4.The workhorse of the previous generation.
Memory144GB HBM3e288GB HBM3e80GB HBM3
Bandwidth—8 TB/s—
Price—~$40–50K (verify)$25–40K
VendorAWSNVIDIANVIDIA
CategoryGPU & AI ChipsGPU & AI ChipsGPU & AI Chips
OfficialAWS ↗NVIDIA ↗NVIDIA ↗

Highlighted rows are where these differ.

Trainium 3

  • EC2 Trn3 only
  • Nearly fully reserved

Best for AWS training.

Full spec sheet →

B300 Blackwell Ultra

  • 8 TB/s memory bandwidth
  • Per-GPU pricing widely misquoted — one source cites rack pricing

Best for frontier training.

Full spec sheet →

H100

The workhorse of the previous generation.

Best for value.

Full spec sheet →