AMD MI300X 192GB vs H100 SXM5 80GB for LLM inference

AMD MI300X 192GB vs H100 SXM5 80GB: verdict

AMD MI300X 192GB vs H100 SXM5 80GB: the short version is that AMD MI300X 192GB wins on VRAM (192GB on the AMD MI300X 192GB vs 80GB on the H100 SXM5 80GB), AMD MI300X 192GB wins on raw memory bandwidth (5300 GB/s on the AMD MI300X 192GB vs 3350 GB/s on the H100 SXM5 80GB, the number that dominates decode throughput), and AMD MI300X 192GB is the cheaper of the two to rent today. Which of those matters more depends entirely on your workload: a large model or long-context deployment cares most about the VRAM number, while a smaller model serving many short requests cares much more about bandwidth and cost per token. When VRAM is tied, bandwidth and price become the primary levers. Neither the AMD MI300X 192GB nor the H100 SXM5 80GB is a universally correct answer — they're built for overlapping but distinct points on the capacity-versus-throughput curve, and the gap between them (in bandwidth and price, and potentially VRAM) is usually the whole reason to pick one over the other rather than defaulting to whichever one you already have access to. The full spec table above covers bf16/fp8 compute throughput, NVLink support, and power draw for both cards, and the models table above shows which currently-popular open-weight models actually fit on each one — the more concrete way to answer 'which GPU do I need' than specs alone.