AI TRAINING
Hopper Data Center H100 SXM Hopper
NVIDIA H100 SXM5
The GPU that powers the AI revolution. The H100 SXM5 delivers 3,958 TFLOPS of FP8 AI performance with 80 GB of HBM3 and 3.35 TB/s of memory bandwidth — engineered for training the world's largest language and multimodal models.
16,896 CUDA Cores
80 GB HBM3
3350 GB/s Bandwidth
700W TDP
MSRP $30,000
Architecture Hopper
GPU Chip GH100
Fabrication TSMC 4N
CUDA Cores 16,896
SM Count 132
Tensor Cores 528
RT Cores 0
Boost Clock 1,980 MHz
Memory Size 80 GB HBM3
Memory Bus 5120-bit
Memory Bandwidth 3350 GB/s
L2 Cache 51 MB
TDP 700 W
Launch Date September 20, 2022
- 4th-Gen Tensor Cores (FP8 / TF32)
- Transformer Engine
- NVLink 4.0 (900 GB/s)
- NVSwitch for DGX H100
- Multi-Instance GPU (MIG)
- Confidential Computing
- SXM5 High-Bandwidth Interconnect
LLM Training
Generative AI
HPC Simulation
Drug Discovery
Climate Modeling