Based on its specifications -- 141GB of HBM3e ECC memory, PCIe Gen 5 x16, dual-slot form factor, NVLink support with up to three bridges, and a 600W power envelope -- this is an NVIDIA H200 Tensor Core GPU, sold through PNY as an NVIDIA board partner. The H200 is a data-centre AI accelerator built for large language model training and inference, HPC simulation, and other memory-bandwidth-hungry workloads; its 141GB of HBM3e is a substantial step up from the earlier H100's 80GB, letting a single card hold much larger models or batch sizes. NVLink bridges allow multiple H200 cards to be pooled with high-bandwidth GPU-to-GPU interconnect inside a server. This is specialised data-centre/AI infrastructure hardware, not a gaming or workstation graphics card, and is typically deployed in purpose-built GPU servers.