H200
NVIDIA's Hopper-generation GPU, a memory-upgraded H100 that serves as the common baseline generation in fleet, confidential-computing, and cross-vendor performance comparisons.
Current numbers
0.5–5 GB/s/GPUDGX H200 SuperPOD shared-storage read guidelines: 0.5, 1, and 5 GB/s per GPU for Good, Better, and Best performance, with 4 GB/s/GPU recommended for computer-vision workloads
~288 GBHBM4 per Rubin GPU at ~22 TB/s; trajectory H100 80 GB → H200 141 GB → B200 180 GB → B300 288 GB → Rubin Ultra 1 TB
~3–12 moH100/H200 GPU lead times through 2025–2026 — bimodal in 2026: spot H100s improving to weeks while H200 and full enterprise systems run 6–12 mo; why GPUs are owner-furnished and on allocation