H100 HGX

NVIDIA H100 & H200 Tensor Core GPUs

The NVIDIA H100 and NVIDIA H200 are the world's most powerful GPUs, based on the latest NVIDIA Hopper architecture, and available in PCIe and SXM form factors.

LAMBDA CLOUD WITH NVIDIA H100 & H200

Lambda powered by NVIDIA H100 & H200

Lambda Reserved Cloud

Reserved cloud clusters from 64 to 32,000 H100 or H200 SXM GPUs

Lambda On-Demand Cloud

On-demand access to 1x H100 PCIe GPUs and 8x HGX H100 instances

1-CLICK CLUSTERS

Introducing 1-Click Clusters

On-demand GPU clusters featuring NVIDIA H100 Tensor Core GPUs with Quantum-2 InfiniBand. No long-term contract required. Self-serve directly from the Lambda Cloud dashboard.
1cc_loop_short (1)
NVIDIA DGX PLATFORM

Lambda offers NVIDIA DGX™ AI supercomputing solutions

Whether creating quality customer experiences, delivering better patient outcomes, or streamlining the supply chain,
enterprises need infrastructure that can deliver AI-powered insights. The NVIDIA DGX platform delivers the
world’s leading solutions for enterprise AI supercomputing at scale.

nvidia dgx h100

NVIDIA DGX™ H100

The fourth generation of the world's most advanced AI system, providing maximum performance.

nvidia dgx superpod

NVIDIA DGX™ SuperPOD

Turnkey, full-stack, industry-leading infrastructure solution for the fastest path to AI innovation at scale.

dgx-cloud-kv-bb460_420-d@2x

NVIDIA DGX™ Cloud

AI-training-as-a-service platform, offering a serverless experience for enterprise developers that is optimized for generative AI.

NVIDIA H200

NVIDIA H200 supercharges generative AI

As the first GPU with HBM3e, H200’s faster, larger memory fuels the acceleration of generative AI and LLMs while advancing scientific computing for HPC workloads. 

Nearly double the GPU memory

The NVIDIA H200 GPU, with 141GB of HBM3e memory, nearly doubles capacity over the prior generation H100. The H200's increased GPU memory capacity allows larger models to be loaded into memory or larger batch sizes for more efficient training of massive LLMs.

Unmatched memory bandwidth

NVIDIA H200’s HBM3e memory bandwidth of 4.8TB/s is 1.4x faster than NVIDIA H100 with HBM3. Increased memory bandwidth and capacity is critical for the growing data sets and model sizes of today’s leading LLMs.

BENCHMARKS
GPU benchmarks

Lambda GPU comparisons

Lambda's GPU benchmarks for deep learning are run on more than a dozen different GPU types in multiple configurations. GPU performance is measured running models for computer vision (CV), natural language processing (NLP), text-to-speech (TTS), and more. Visit our benchmarks page to get started.

TECH SPECS
  H100 HGX H100 HGX
H100_PCIe
Model H200 SXM H100 SXM H100 PCIe
GPU memory 141GB HBM3e at 4.8 TB/s 80GB HBM3 at 3.35 TB/s 80GB HBM2e at 2 TB/s
Form Factor 8-GPU SXM5 8-GPU SXM5 PCIe Dual-slot
Interconnect 900GB/s NVLink 900GB/s NVLink 600GB/s NVLink
Cloud Options Lambda Reserved Cloud Lambda On-Demand Cloud
Lambda Reserved Cloud
NVIDIA DGX Cloud
Lambda On-Demand Cloud
Server options NVIDIA DGX Server
NVIDIA DGX SuperPOD
Lambda Scalar
NVIDIA AI Enterprise Add-on Add-on Included
Now available

Lambda On-Demand Cloud powered by NVIDIA H100 GPUs

On-demand HGX H100 systems with 8x NVIDIA H100 SXM GPUs are now available on Lambda Cloud for only $2.59/hr/GPU. With H100 SXM you get: 

  • More flexibility for users looking for more compute power to build and fine-tune generative AI models
  • Enhanced scalability
  • High-bandwidth GPU-to-GPU communication
  • Optimal performance density

Lambda Cloud also has 1x NVIDIA H100 PCIe GPU instances at just $1.99/hr/GPU for smaller experiments.

RESOURCES

Resources for deep learning

Explore Lambda's deep learning materials including blog, technical documentation, research and more. We've curated a diverse set of resources just for ML and AI professionals to help you on your journey.