GPU-192xCPU-1920GB-8xB200 by UpCloud

A high-density parallel computing platform featuring eight Blackwell GPUs and nearly two terabytes of system memory for massive AI workloads.

The UpCloud GPU-192xCPU-1920GB-8xB200 is a high-density GPU-accelerated server instance designed for massive parallel computing workloads. Operating on the KVM hypervisor with an x86_64 architecture, the system features 192 shared vCPUs and 1920.0 GB of system memory, providing 10.0 GB of RAM per core. The primary hardware highlight is the inclusion of eight NVIDIA Blackwell B200 GPUs, delivering a combined total of 1536 GB of VRAM. This server does not include local storage, requiring external storage attachments, and includes one complimentary public IPv4 address. Given its massive GPU memory and high system RAM capacity, this instance is engineered for intensive workloads such as large-scale artificial intelligence model training, large language model inference, and complex data science computations where GPU-bound processing is the primary bottleneck.