Live
AI hardware

NVIDIA H800 SXM5

FP16 throughput
9.9×10¹⁴ FLOP/s
Memory
74.5 GB
TDP
700 W

NVIDIA H800 SXM5 is an AI accelerator designed by NVIDIA (GPU), released in March 2023. It is built on a 5 nm process at TSMC, packing 80B transistors.

It delivers 9.9×10¹⁴ FLOP/s of tensor FP16/BF16 throughput. It carries 74.5 GB of memory at 3.1 TB/s of bandwidth. Its thermal design power is 700 W.

Full record
Manufacturer
NVIDIA
Type
GPU
Release date
Mar 21, 2023
Tensor FP16/BF16
9.9×10¹⁴ FLOP/s
FP32
5.9×10¹³ FLOP/s
FP64
3×10¹³ FLOP/s
INT8
2×10¹⁵ OP/s
Memory
74.5 GB
Memory bandwidth
3.1 TB
TDP
700 W
Process node
5 nm
Transistors (millions)
80,000
Die size (mm2)
814
Foundry
TSMC
Tensor cores
456
More from NVIDIA
SourceEpoch AI, 'Machine Learning Hardware'. Published online at epoch.ai. Retrieved 2026-07-29 from https://epoch.ai/data/machine-learning-hardware. Licensed under CC BY 4.0.
← All ai hardware