NVIDIA H200 GPU
$1.50 - $2.25
/hr
The NVIDIA H200 Tensor Core GPU, based on NVIDIA’s Hopper architecture, is designed for large-scale AI workloads that benefit from increased memory capacity and bandwidth, including LLM training, fine-tuning, and high-throughput inference.
Compared to H100, H200 significantly expands available HBM memory and bandwidth, making it particularly well suited for memory-bound workloads such as large transformer models, long-context inference, and data-intensive training pipelines.
H200 GPUs are commonly deployed as reserved capacity by AI startups, research organizations, and enterprises running sustained training or production inference workloads.
NVIDIA H200 SPECIFICATIONS
ARCHITECTURE
NVIDIA Hopper
MEMORY
141 GB HBM3e
MEMORY BANDWIDTH
Up to ~4.8 TB/s
PRECISION SUPPORT
FP8 / FP16 / BF16 / TF32
INTERCONNECT
Up to 900 GB/s NVLink
FORM FACTORS
SXM5
PRIMARY USE CASES
Large language model training, Memory-bound fine-tuning, High-throughput inference
TDP
700W (SXM)
WHAT IS H200 USED FOR?
H200 pricing varies based on region, provider, system configuration, networking, and availability. As a newer generation compared to H100, access to H200 capacity is often tied to longer-term commitments and reserved capacity agreements.
Live H100 availability
Regional price differences
Configuration comparisons
Flexible deployment options
Reserved H200 capacity across cloud, neocloud, and independent providers
Neocloud providers with optimized AI infrastructure
Contract-based allocations with defined terms and guaranteed availability
Single-node and multi-node cluster reservations
Commitments aligned with sustained training or production inference workloads
"Compute Exchange acts as a broker and marketplace layer, helping buyers match workload needs to the right supply — without forcing architectural changes."
Global Network
Access verified providers across North America, Europe, and Asia Pacific, offering reserved capacity in multiple regions and configurations.
Why Buy H100 Through Compute Exchange
The NVIDIA GB200 Grace Blackwell Superchip connects two NVIDIA B200 Tensor Core GPUs to the NVIDIA Grace CPU over a high-speed NVLink-C2C interconnect.