Reserve NVIDIA HGX H200 capacity on CoreWeave today

Enterprise-ready infrastructure. Purpose-built for speed, memory, and efficiency.

Why choose CoreWeave for H200s?

With expanded HBM3e memory and unmatched bandwidth, CoreWeave Cloud enables you to use H200s to train trillion-parameter models, accelerate inference for next-gen AI workloads at scale, and optimize efficiency across the full AI lifecycle. Get AI innovations to market smarter and faster.

• Next-level inference speed
  CoreWeave's NVIDIA H200 GPU instances deliver 33,000 TPS on the Llama 2 70B model, marking a 40% improvement over our closest competition.
• Unmatched GPU throughput
  CoreWeave delivers deep observability, bare-metal performance isolation, and proactive infrastructure management, maintaining up to 96% goodput for large-scale training workloads.
• Transparent, predictable pricing
  Straightforward models with no hidden fees, enabling finance teams to accurately forecast and control spend as AI initiatives scale.