Skip to main content

NVIDIA® H200 NVL

The NVIDIA H200 NVL brings the power of HBM3e memory to a PCIe NVLink form factor, enabling high-performance AI workloads in standard server configurations.

Availability

Available on Rackrr through our marketplace providers. Check the platform for real-time pricing and availability.

Specifications

ComponentDetails
GPU ModelNVIDIA® H200 NVL
ArchitectureNVIDIA Hopper
CUDA Cores16,896
Tensor Cores528 (4th Gen)
VRAM141 GB
VRAM TypeHBM3e
Memory Bandwidth4.8 TB/s
TDP (Thermal Design Power)600W
Form FactorPCIe NVL
InterconnectNVLink Bridge (600 GB/s)
FP8 Performance3,958 TFLOPS
FP16 Performance1,979 TFLOPS
Deep Learning FrameworksPyTorch, TensorFlow, JAX

Use Cases

  • AI inference at scale
  • Generative AI workloads
  • Retrieval-augmented generation (RAG)
  • High-performance computing (HPC)