Shop
Showing 604–612 of 755 results
-
NVIDIA DGX Spark Personal AI Supercomputer
Read more
The NVIDIA DGX Spark is a compact personal AI supercomputer built around the NVIDIA GB10 Grace Blackwell Superchip, combining a 20-core Arm CPU, Blackwell GPU architecture, 128GB of coherent unified LPDDR5x memory, high-speed ConnectX-7 networking, and NVIDIA’s full AI software ecosystem in a small desktop system.
With up to 1 PFLOP of FP4 AI performance, 6,144 CUDA cores, fifth-generation Tensor Cores, and 128GB of unified CPU/GPU memory, DGX Spark is designed to bring advanced AI development, inference, fine-tuning, data science, and generative AI workloads directly to the desktop.
The system supports local AI model development and inference for models of up to approximately 200 billion parameters, making it particularly attractive for AI developers, researchers, data scientists, engineers, and organizations that want local AI computing without depending entirely on cloud infrastructure.
Quick View -
NVIDIA GB200 NVL72 – Grace Blackwell Rack-Scale AI Supercomputer
Read more
The NVIDIA GB200 NVL72 is a rack-scale accelerated computing platform built on the NVIDIA Grace Blackwell architecture, combining 36 NVIDIA Grace CPUs and 72 NVIDIA Blackwell GPUs into a single high-performance NVLink domain. Designed for trillion-parameter AI models, generative AI, large-scale inference, model training, and HPC, the GB200 NVL72 delivers massive GPU memory capacity, extremely high memory bandwidth, and ultra-high-speed GPU-to-GPU communication.
The system provides up to 13.4TB of HBM3e GPU memory, 576TB/s aggregate GPU memory bandwidth, and 130TB/s of NVLink bandwidth, creating a tightly coupled 72-GPU computing domain optimized for large-scale AI workloads.
Quick View -
NVIDIA GH200 Grace Hopper Superchip – 96GB HBM3 AI & HPC Accelerator
Read more
The NVIDIA GH200 Grace Hopper Superchip combines a 72-core NVIDIA Grace Arm CPU with an NVIDIA Hopper GPU featuring 96GB HBM3 in a tightly integrated heterogeneous computing platform. Connected through 900GB/s coherent NVLink-C2C, GH200 provides a unified CPU-GPU memory architecture designed for large-scale AI, generative AI, high-performance computing, scientific computing, and data-intensive workloads.
Quick View -
NVIDIA H100 PCIe 80GB – Hopper Tensor Core Data Center GPU
Read more
The NVIDIA H100 PCIe 80GB is a high-performance Hopper-generation Tensor Core data-center GPU designed for artificial intelligence, machine learning, high-performance computing (HPC), scientific computing, and accelerated data analytics. It features 80GB HBM2e memory, PCIe Gen5 x16 connectivity, approximately 2TB/s memory bandwidth, and support for Multi-Instance GPU (MIG) technology with up to seven GPU instances.
Quick View -
NVIDIA H100 SXM5 80GB – Hopper Tensor Core Data Center GPU
Read more
The NVIDIA H100 SXM5 80GB is a flagship Hopper-generation data-center GPU accelerator designed for large-scale artificial intelligence, machine learning, generative AI, high-performance computing (HPC), and scientific workloads. It combines 80GB HBM3 memory, 3.35 TB/s memory bandwidth, 16,896 CUDA cores, fourth-generation Tensor Cores, and 900 GB/s NVIDIA NVLink connectivity for highly scalable multi-GPU computing.
Quick View -
NVIDIA H200 PCIe 141GB – Hopper Tensor Core Data Center GPU
Read moreThe NVIDIA H200 PCIe 141GB is a high-performance Hopper-generation data-center GPU accelerator engineered for artificial intelligence, generative AI, large language models, high-performance computing (HPC), scientific computing, and accelerated data analytics. It features 141GB of HBM3e memory, up to 4.8TB/s memory bandwidth, 16,896 CUDA cores, fourth-generation Tensor Cores, and a PCIe Gen5 x16 host interface.
Quick View -
NVIDIA H200 SXM 141GB – Hopper Tensor Core Data Center GPU
Read more
The NVIDIA H200 SXM 141GB is a high-performance Hopper-generation data-center GPU accelerator designed for large-scale artificial intelligence, generative AI, high-performance computing (HPC), and scientific workloads. It features 141GB of HBM3e memory, up to 4.8TB/s memory bandwidth, 16,896 CUDA cores, fourth-generation Tensor Cores, and high-speed NVIDIA NVLink connectivity for scalable multi-GPU computing.
Quick View -
NVIDIA HGX B200 NVL8
Read more
The NVIDIA HGX B200 NVL8 is a high-density 8-GPU AI and high-performance computing platform built around eight NVIDIA Blackwell B200 Tensor Core GPUs connected through a high-bandwidth NVIDIA NVLink/NVSwitch fabric. With 1.44TB of aggregate HBM3e GPU memory, up to 64TB/s of aggregate GPU memory bandwidth, and up to 1.8TB/s GPU-to-GPU bandwidth, HGX B200 NVL8 is engineered for large-scale generative AI, deep learning, LLM training and inference, scientific computing, and enterprise AI infrastructure.
Quick View -
NVIDIA L4 24GB – Ada Lovelace Data Center GPU
Read more
The NVIDIA L4 24GB is a compact, energy-efficient data-center GPU designed for AI inference, video processing, graphics virtualization, professional visualization, and cloud computing.
Built on the NVIDIA Ada Lovelace architecture, the L4 features 7,424 CUDA cores, 58 fourth-generation Tensor Cores, 24 third-generation RT Cores, and 24GB of ECC-protected GDDR6 memory. Its 300GB/s memory bandwidth and exceptionally low 72W maximum power consumption make it well suited to high-density servers and edge deployments.
Quick View
Loading... All products loaded