Jake Tech Lab

NVIDIA A100 vs NVIDIA H200

NVIDIA H200 comes out ahead of NVIDIA A100 on 7 of the 8 things that decide this, going by what each maker publishes. The one that matters for you may be one the leader loses.

NVIDIA A100
NVIDIA A100
6.0
Fair
1 of 8 rounds
vs
NVIDIA H200Winner
NVIDIA H200
6.5
Fair
7 of 8 rounds

The verdict

NVIDIA H200 takes it, 7–1 on rounds.

On the published figures, NVIDIA H200 takes 7 of 8 rounds. Treat that as a starting point and check the current figures on both pages, because prices and plans move.

Choose NVIDIA A100 if…

  • • You want NVIDIA A100's edge on how much power it draws

Choose NVIDIA H200 if…

  • • You want NVIDIA H200's edge on how much memory it has and how fast its memory is

Round by round

  1. Round 1How much memory it hasNVIDIA H200 wins
    NVIDIA A1005/10

    80 GB HBM2e

    gpu memory

    • 80 GB HBM2e, against 141 GB HBM3e

    Source: nvidia.com ↗

    NVIDIA H2008/10

    141 GB HBM3e

    gpu memory

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  2. Round 2How fast its memory isNVIDIA H200 wins
    NVIDIA A1005/10

    2,039 GB/s

    memory bandwidth

    • 2,039 GB/s, against 4.8 TB/s

    Source: nvidia.com ↗

    NVIDIA H2008/10

    4.8 TB/s

    memory bandwidth

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  3. Round 3How fast it runs FP16 AI mathNVIDIA H200 wins
    NVIDIA A1005/10

    624 TFLOPS

    fp16 tensor core, with sparsity

    • 624 TFLOPS, against 1,979 TFLOPS

    Source: nvidia.com ↗

    NVIDIA H2008/10

    1,979 TFLOPS

    fp16 tensor core, with sparsity

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  4. Round 4How much power it drawsNVIDIA A100 wins
    NVIDIA A1008/10

    400 W

    max tdp (sxm)

    Source: nvidia.com ↗

    NVIDIA H2005/10

    700 W

    max tdp

    • 700 W, against 400 W

    Source: nvidia.com ↗

    NVIDIA A100 wins this round on what the company itself publishes.

  5. Round 5How fast chips talk to each otherNVIDIA H200 wins
    NVIDIA A1005/10

    600 GB/s NVLink

    nvlink gpu-to-gpu

    • 600 GB/s NVLink, against 900 GB/s NVLink

    Source: nvidia.com ↗

    NVIDIA H2008/10

    900 GB/s NVLink

    nvlink gpu-to-gpu

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  6. Round 6How it handles scientific FP64 mathNVIDIA H200 wins
    NVIDIA A1005/10

    19.5 TFLOPS

    fp64 tensor core

    • 19.5 TFLOPS, against 67 TFLOPS

    Source: nvidia.com ↗

    NVIDIA H2008/10

    67 TFLOPS

    fp64 tensor core

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  7. Round 7Which memory generation it usesNVIDIA H200 wins
    NVIDIA A1005/10

    HBM2e

    memory type

    • HBM2e, against HBM3e

    Source: nvidia.com ↗

    NVIDIA H2008/10

    HBM3e

    memory type

    Source: nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

  8. Round 8How new it isNVIDIA H200 wins
    NVIDIA A1005/10

    2020

    year announced

    • 2020, against 2023

    Source: nvidianews.nvidia.com ↗

    NVIDIA H2008/10

    2023

    year announced

    Source: nvidianews.nvidia.com ↗

    NVIDIA H200 wins this round on what the company itself publishes.

NVIDIA A100NVIDIA A100
NVIDIA H200NVIDIA H200

Side by side

Overall score
6.0
6.5
Best
Price
See price
See price
How fast its memory is
6.1
6.0
How much memory it has
6.3
6.0
How much power it draws
6.9
7.4
How fast it runs FP16 AI math
5.8
6.3
How new it is
5.4
6.3
Head-to-head findings
How fast its memory is
2,039 GB/s
4.8 TB/s
How much memory it has
80 GB HBM2e
141 GB HBM3e
How much power it draws
400 W
700 W
How fast it runs FP16 AI math
312 TFLOPS
1,979 TFLOPS
How new it is
2020
2023
Which memory generation it uses
HBM2e
HBM3e
How fast it runs FP8 AI math
—
3,958 TFLOPS
How it handles scientific FP64 math
19.5 TFLOPS
67 TFLOPS
Record
Head-to-head record
2–5–1
2–4
Head-to-heads won
29 %
33 %
Rounds won
31 %
43 %
Pros
      Cons
      • 80 GB HBM2e, against 128 GB HBM2E
      • 2,039 GB/s, against 3.7 TB/s
      • 312 TFLOPS, against 1,678 TFLOPS
      • 2020, against September 24, 2024
      • 141 GB HBM3e, against 186 GB HBM3E
      • 4.8 TB/s, against 8 TB/s
      • 3,958 TFLOPS, against 10 PFLOPS
      • 1,979 TFLOPS, against 5 PFLOPS
      Add a third option to this comparison →

      More AI accelerators match-ups