{ "version": "1", "updated": "2026-10-19", "data_version": "2026-10-19", "retrieved": "2026-09-26", "note": "Manufacturer ceilings transcribed from NVIDIA's published pages and documents on the retrieval date. NVIDIA does not publish power draw per MIG slice, and neither public benchmark on the index measures a sliced card. The '+me' media-extension variants of the 1g profiles are omitted.", "tdp": [ { "gpu": "A100 80GB", "form": "SXM", "max_tdp_w": 400, "tdp_label": "400 W", "memory": "80 GB HBM2e", "source": "https://www.nvidia.com/en-us/data-center/a100/" }, { "gpu": "A100 80GB", "form": "PCIe", "max_tdp_w": 300, "tdp_label": "300 W", "memory": "80 GB HBM2e", "source": "https://www.nvidia.com/en-us/data-center/a100/" }, { "gpu": "H100", "form": "SXM", "max_tdp_w": 700, "tdp_label": "up to 700 W", "memory": "80 GB HBM3", "source": "https://www.nvidia.com/en-us/data-center/h100/" }, { "gpu": "H100", "form": "NVL", "max_tdp_w": 400, "tdp_label": "350 to 400 W", "memory": "94 GB HBM3", "source": "https://www.nvidia.com/en-us/data-center/h100/" }, { "gpu": "H200", "form": "SXM", "max_tdp_w": 700, "tdp_label": "up to 700 W", "memory": "141 GB HBM3e", "source": "https://www.nvidia.com/en-us/data-center/h200/" }, { "gpu": "H200", "form": "NVL", "max_tdp_w": 600, "tdp_label": "up to 600 W", "memory": "141 GB HBM3e", "source": "https://www.nvidia.com/en-us/data-center/h200/" }, { "gpu": "B200", "form": "HGX, SXM6", "max_tdp_w": 1000, "tdp_label": "configurable up to 1,000 W", "memory": "180 GB HBM3e", "source": "https://images.nvidia.com/aem-dam/Solutions/documents/HGX-B200-PCF-Summary.pdf" } ], "systems": [ { "system": "DGX B200", "gpus": "8 x B200", "max_power_kw": 14.3, "label": "~14.3 kW max", "source": "https://www.nvidia.com/en-us/data-center/dgx-b200/" } ], "mig_profiles": [ { "gpu": "A100 80GB", "profile": "1g.10gb", "memory_fraction": "1/8", "sm_fraction": "1/7", "instances": 7 }, { "gpu": "A100 80GB", "profile": "1g.20gb", "memory_fraction": "1/4", "sm_fraction": "1/7", "instances": 4 }, { "gpu": "A100 80GB", "profile": "2g.20gb", "memory_fraction": "2/8", "sm_fraction": "2/7", "instances": 3 }, { "gpu": "A100 80GB", "profile": "3g.40gb", "memory_fraction": "4/8", "sm_fraction": "3/7", "instances": 2 }, { "gpu": "A100 80GB", "profile": "4g.40gb", "memory_fraction": "4/8", "sm_fraction": "4/7", "instances": 1 }, { "gpu": "A100 80GB", "profile": "7g.80gb", "memory_fraction": "full", "sm_fraction": "7/7", "instances": 1 }, { "gpu": "H100 80GB", "profile": "1g.10gb", "memory_fraction": "1/8", "sm_fraction": "1/7", "instances": 7 }, { "gpu": "H100 80GB", "profile": "1g.20gb", "memory_fraction": "1/4", "sm_fraction": "1/7", "instances": 4 }, { "gpu": "H100 80GB", "profile": "2g.20gb", "memory_fraction": "2/8", "sm_fraction": "2/7", "instances": 3 }, { "gpu": "H100 80GB", "profile": "3g.40gb", "memory_fraction": "4/8", "sm_fraction": "3/7", "instances": 2 }, { "gpu": "H100 80GB", "profile": "4g.40gb", "memory_fraction": "4/8", "sm_fraction": "4/7", "instances": 1 }, { "gpu": "H100 80GB", "profile": "7g.80gb", "memory_fraction": "full", "sm_fraction": "7/7", "instances": 1 }, { "gpu": "H200 141GB", "profile": "1g.18gb", "memory_fraction": "1/8", "sm_fraction": "1/7", "instances": 7 }, { "gpu": "H200 141GB", "profile": "1g.35gb", "memory_fraction": "1/4", "sm_fraction": "1/7", "instances": 4 }, { "gpu": "H200 141GB", "profile": "2g.35gb", "memory_fraction": "2/8", "sm_fraction": "2/7", "instances": 3 }, { "gpu": "H200 141GB", "profile": "3g.71gb", "memory_fraction": "4/8", "sm_fraction": "3/7", "instances": 2 }, { "gpu": "H200 141GB", "profile": "4g.71gb", "memory_fraction": "4/8", "sm_fraction": "4/7", "instances": 1 }, { "gpu": "H200 141GB", "profile": "7g.141gb", "memory_fraction": "full", "sm_fraction": "7/7", "instances": 1 }, { "gpu": "B200 180GB", "profile": "1g.23gb", "memory_fraction": "1/8", "sm_fraction": "1/7", "instances": 7 }, { "gpu": "B200 180GB", "profile": "1g.45gb", "memory_fraction": "1/4", "sm_fraction": "1/7", "instances": 4 }, { "gpu": "B200 180GB", "profile": "2g.45gb", "memory_fraction": "2/8", "sm_fraction": "2/7", "instances": 3 }, { "gpu": "B200 180GB", "profile": "3g.90gb", "memory_fraction": "4/8", "sm_fraction": "3/7", "instances": 2 }, { "gpu": "B200 180GB", "profile": "4g.90gb", "memory_fraction": "4/8", "sm_fraction": "4/7", "instances": 1 }, { "gpu": "B200 180GB", "profile": "7g.180gb", "memory_fraction": "full", "sm_fraction": "7/7", "instances": 1 } ], "embodied": { "item": "NVIDIA HGX B200 baseboard, eight GPUs", "kg_co2e_cradle_to_gate": 2274, "source": "https://images.nvidia.com/aem-dam/Solutions/documents/HGX-B200-PCF-Summary.pdf", "published": "2025-07" }, "sources": [ { "name": "NVIDIA Multi-Instance GPU User Guide, Supported MIG Profiles", "url": "https://docs.nvidia.com/datacenter/tesla/mig-user-guide/supported-mig-profiles.html", "version": "last updated 2026-09-11", "retrieved": "2026-09-26", "used_for": "mig_profiles" }, { "name": "NVIDIA product specifications: A100, H100, H200, DGX B200", "url": "https://www.nvidia.com/en-us/data-center/h100/", "version": "as published on the retrieval date", "retrieved": "2026-09-26", "used_for": "tdp (A100, H100, H200), systems" }, { "name": "NVIDIA, Product Carbon Footprint Summary for NVIDIA HGX B200", "url": "https://images.nvidia.com/aem-dam/Solutions/documents/HGX-B200-PCF-Summary.pdf", "version": "July 2025", "retrieved": "2026-09-26", "used_for": "tdp (B200), embodied" } ] }