10× NVIDIA H100 80GB Tensor Core GPU — PCIe 5.0 x16 | Hopper Enterprise AI / HPC

Item description from the seller

Key Specifications & Features

Quantity: 10× NVIDIA H100 GPUs (PCIe 5.0 x16 interface)

Memory: 80 GB HBM2e per GPU (total 800 GB)

Architecture: NVIDIA Hopper (H100 Tensor Core)

Performance: Up to 2,000 TFLOPS (FP8), 32 TFLOPS (FP64) aggregate theoretical performance

Power: 350W per GPU (PCIe variant)

Connectivity: PCIe Gen 5.0 x16 — compatible with most enterprise-grade motherboards and workstations

Best For: AI training, LLMs, HPC clusters, data analytics, inference acceleration

Performance & Use Cases

Built for large-scale AI training and inference workloads

Enables multi-GPU scaling and distributed deep learning systems

Ideal for model fine-tuning, simulation workloads, and next-generation data center applications

Supported by NVIDIA CUDA, cuDNN, TensorRT, and NVLink (via bridge kits, sold separately)

What You’ll Receive

10× Brand new NVIDIA H100 80GB GPUs (PCIe 5.0 x16)

Original factory packaging and protection (anti-static, foam inserts)

OEM serials intact for authenticity verification

Optional professional installation and configuration available upon request

Condition: New, factory sealed, unopened — in original NVIDIA OEM packaging