New GPU Servers: NVIDIA L4, H100, and B200

We are expanding our GPU Server lineup with three new GPU options — NVIDIA L4, H100, and B200 — joining the existing NVIDIA L40S plans.

NVIDIA L4 is well-suited for image generation, speech-to-text, and basic inference workloads. Plans are available with 1, 2, or 3 GPUs and up to 32 CPU cores and 384 GB of RAM.

NVIDIA H100 targets high-traffic inference, large-scale batch processing, and model training. Multi-GPU configurations of 2, 4, and 8 GPUs are available, scaling up to 96 CPU cores and 1920 GB of RAM. All H100 plans include NVLink, providing 900 GB/s of bidirectional GPU-to-GPU bandwidth for efficient distributed workloads.

NVIDIA B200 is designed for the most demanding AI workloads, including trillion-parameter model inference and real-time complex model execution. Multi-GPU configurations with NVLink are available, offering 1.8 TB/s of GPU-to-GPU bandwidth.

All new plans follow the same billing model as existing GPU Servers — you are only charged when the server is powered on. See the GPU Server configurations page for the full list of available plans.