Manual · GPU

H200: Hopper with a bigger tank

Same compute as H100, but 141GB of HBM3e. For inference, memory is the bottleneck, and H200 is the memory play.

H200 is the answer to a question nobody asked in 2022 but everyone needed by 2024: what if the compute stayed the same but the memory got huge? For inference, that's often the whole ballgame.

What changed from H100

The numbers

Why it matters for inference

Decode is memory-bound. The GPU streams weights once per token, and the bandwidth determines the token rate. More bandwidth = faster decode. More capacity = bigger KV cache, longer context, more concurrent requests. H200 is the GPU you pick when you want maximum tokens per second per dollar on decode-heavy workloads.

It's also the GPU that made 70B models comfortable: 70B in FP8 is ~70GB, and with 141GB you have 70GB left for KV cache and activations. That's a lot of concurrent long-context sessions on one GPU.

Sources

Back to the manual