Unrivalled
Performance in..
Generative AI
Up to 2x the speed of the NVIDIA H100 for model training and inference.
LLM Inference
Delivers 2x the performance of the NVIDIA H100 for LLMs like Llama2 70 B.
High-Performance Computing (HPC)
Offers 110x faster MILC performance than dual x86 CPUs, perfect for memory-heavy tasks.
AI & Machine Learning (ML)
3,958 TFLOPS of FP8 compute supercharge AI model training for lightning-fast results.
Key Features of
NVIDIA H200 SXM
Scalable Configuration
Deploy the NVIDIA H200-141G-SXM5x8 configuration, which combines eight powerful H200 SXM GPUs, ideal for enterprise-scale generative AI, model training and memory-intensive simulations.
High-Speed Networking
Experience high-speed networking of up to 350 Gbps for low-latency and high-throughput workloads. Ideal for ensuring fast data transfer and minimal delays for large-scale workloads.
Massive Ephemeral
NVMe Storage
Get 32 TB of ultra-fast ephemeral storage, perfect for high-speed data processing, temporary caching and handling large training datasets during intensive AI and HPC workloads.
Huge RAM Capacity
Enjoy 1920 GB RAM for smooth operation of large AI models, dataset handling and compute-heavy applications without running into memory bottlenecks.
Snapshot Support for
Fast Recovery
Take system snapshots to capture your NVIDIA H200 SXM VM’s exact state, including config and bootable volumes. Restore environments quickly and safely for versioning, rollback or disaster recovery.
Bootable Volume
for Persistent OS State
Every NVIDIA H200 SXM VM comes with a dedicated 100 GB bootable volume, holding OS files and key settings to run the VM’s operating system.
NVIDIA H200 SXM
Starts from $2.45/hour
