Nvidia
Read More
Product Filter
Showing 25–36 of 63 results
-
NVIDIA RTX PRO 6000 Blackwell Server Edition 900-2G153-0000-000
$Quote1 pcs(MOQ)
The NVIDIA RTX PRO 6000 Blackwell Server Edition (900-2G153-0000-000) is a high-performance professional GPU designed for data centers and advanced computing workloads. Powered by the Blackwell architecture, it delivers exceptional AI, graphics, and compute performance with enhanced energy efficiency. Key features include massive CUDA core counts, next-gen RT and Tensor cores for real-time ray tracing and AI acceleration, and large memory capacity optimized for complex simulations and rendering tasks. Its server-grade design ensures reliability, scalability, and seamless integration into enterprise environments, making it ideal for professionals in AI research, 3D rendering, and high-end visualization.
Bulk Order Discounts Available -
NVIDIA H200 Tensor Core GPU,141 GB HBM3e, NVLink, PCIe 5.0, 700 W
$Quote1 Piece(MOQ)
The NVIDIA H200 Tensor Core GPU is a high-performance accelerator designed for AI, HPC, and data analytics workloads. It features 141 GB of ultra-fast HBM3e memory, delivering exceptional bandwidth for large-scale models and datasets. With support for NVLink and PCIe 5.0, it ensures high-speed interconnectivity and scalability across systems. Operating at 700 W, the H200 offers unmatched computational power, making it ideal for demanding generative AI and large language model applications. Its cutting-edge architecture and memory capacity position it as a top-tier solution for next-generation AI infrastructure.
Bulk Order Discounts Available -
NVIDIA A100 80G SXM4
$Quote1 Piece(MOQ)
The NVIDIA A100 80G SXM4 is a high-performance GPU designed for data centers, delivering exceptional acceleration for AI, machine learning, and high-performance computing (HPC) workloads. Featuring 80GB of high-bandwidth HBM2e memory and third-generation Tensor Cores, it enables massive model training, real-time inference, and advanced analytics. Built on the NVIDIA Ampere architecture, it supports multi-instance GPU (MIG) technology for optimal resource utilization and scalability. Its SXM4 form factor ensures maximum throughput and energy efficiency, making it ideal for demanding enterprise and research applications.
Bulk Order Discounts Available -
NVIDIA 900-9D3B6-00CV-AA0（B3220）P-Series BlueField – 3 NDR200Gb/s dual port QSFP112 InfiniBand/Ethernet DPU
$Quote5 pcs(MOQ)
The NVIDIA 900-9D3B6-00CV-AA0 (B3220) P-Series BlueField-3 DPU is a high-performance data processing unit designed for next-generation data centers. It features dual-port QSFP112 interfaces supporting NDR 200Gb/s InfiniBand and Ethernet, delivering ultra-low latency and high throughput. Built on the powerful BlueField-3 architecture, it integrates 16 Arm cores, hardware accelerators for security, storage, and networking, and supports offloading of infrastructure workloads from CPUs. Key benefits include enhanced performance, improved security isolation, and reduced total cost of ownership. Its unique selling point is the combination of advanced programmability and acceleration, making it ideal for AI, HPC, and cloud-native environments.
Bulk Order Discounts Available -
NVIDIA T1000 8GB GDDR6 PCIe3.0 900-5G172-2270-000 Professional Cards
$Quote1 pcs(MOQ)
The NVIDIA T1000 8GB GDDR6 is a compact, professional-grade graphics card designed for demanding visual computing tasks. Featuring 896 CUDA cores and 8GB of high-speed GDDR6 memory, it delivers powerful performance for 2D and 3D workflows, CAD, and content creation. Its low-profile, single-slot design ensures compatibility with a wide range of workstations, while support for up to four 4K displays via mini DisplayPort 1.4 enhances productivity. Built on the Turing architecture and optimized for stability and efficiency, the T1000 offers a perfect balance of performance, size, and reliability for professionals.
Bulk Order Discounts Available -
NVIDIA Quadro T1000 8G – Professional Graphics Card 900-5G172-2570-000
$Quote1 pcs(MOQ)
The NVIDIA Quadro T1000 8G is a professional-grade graphics card designed for creative and technical professionals. Featuring 8GB of GDDR6 memory and powered by NVIDIA’s Turing architecture, it delivers exceptional performance for CAD, DCC, and visualization workflows. With 896 CUDA cores, it supports real-time rendering, complex 3D modeling, and multi-display setups via four mini DisplayPort 1.4 outputs. Its compact, low-profile form factor ensures compatibility with a wide range of workstations. The Quadro T1000 8G offers a balance of power, efficiency, and reliability, making it ideal for professionals seeking high performance in space-constrained environments.
Bulk Order Discounts Available -
NVIDIA L20 48GB GPU
$Quote1 Piece(MOQ)
The NVIDIA L20 48GB GPU is a high-performance accelerator designed for AI inference, data analytics, and virtual desktop workloads. Featuring 48GB of high-bandwidth memory, it delivers exceptional throughput for large models and datasets. Built on the NVIDIA Ada Lovelace architecture, it offers advanced tensor cores, improved energy efficiency, and support for FP8 precision, enabling faster and more cost-effective AI deployment. With multi-instance GPU (MIG) support and enterprise-grade reliability, the L20 is ideal for data centers seeking scalable, secure, and efficient AI infrastructure.
Bulk Order Discounts Available -
NVIDIA/Mellanox MCX75310AAS-HEAT ConnectX-7 InfiniBand Adapter Card, NDR200 200Gb/s, Single-Port OSFP, PCIe 4.0/5.0 x 16, Tall Bracket
$Quote1 Piece(MOQ)
The NVIDIA/Mellanox MCX75310AAS-HEAT ConnectX-7 is a high-performance InfiniBand adapter card designed for data-intensive and latency-sensitive applications. It delivers blazing-fast 200Gb/s NDR200 throughput via a single OSFP port and supports PCIe 4.0/5.0 x16 interfaces for maximum bandwidth. Engineered for next-generation data centers, HPC, and AI workloads, it offers ultra-low latency, advanced offloads, and hardware acceleration for RDMA, GPUDirect, and in-network computing. The tall bracket form factor ensures compatibility with standard server configurations. Its cutting-edge performance and scalability make it ideal for demanding enterprise and scientific computing environments.
Bulk Order Discounts Available -
NVIDIA DGX Spark Grace Blackwell GB10 Superchip Desktop AI Supercomputer NVIDIA DGX Spark GB10
$Quote5 pcs(MOQ)
The NVIDIA DGX Spark Grace Blackwell GB10 Superchip Desktop AI Supercomputer is a cutting-edge AI system designed for extreme performance in desktop form. Powered by the NVIDIA Grace Blackwell GB10 Superchip, it delivers unprecedented compute power, combining advanced CPU and GPU architectures for seamless AI, data analytics, and HPC workloads. Key features include ultra-high memory bandwidth, energy-efficient performance, and native support for large-scale generative AI models. Its compact design brings data center-level capabilities to the desktop, making it ideal for researchers, developers, and enterprises seeking accelerated AI innovation. Unique selling points include its integration of Grace CPU and Blackwell GPU on a single superchip, optimized for AI inference and training with minimal latency and maximum throughput.
Bulk Order Discounts Available -
NVIDIA ConnectX-7 200GbE/NDR200 IB Dual-Port QSFP112 OCP 3.0 SFF MCX753436MS-HEAB Adapter Card
$1534.68/pcs1 Piece(MOQ)
The NVIDIA ConnectX-7 MCX753436MS-HEAB is a high-performance network adapter card designed for data centers and high-throughput computing environments. Featuring dual-port QSFP112 interfaces, it supports up to 200GbE Ethernet and NDR200 InfiniBand connectivity, delivering ultra-low latency and exceptional bandwidth. Built on the OCP 3.0 SFF form factor, it ensures seamless integration into modern server infrastructures. Key features include advanced offloads for RDMA, GPUDirect, and hardware acceleration for networking, storage, and security workloads. Its unique selling points are unmatched speed, scalability, and efficiency, making it ideal for AI, HPC, and cloud-scale deployments.
Bulk Order Discounts Available -
NVIDIA MMA4Z00-NS Broadcom VCSEL Compatible 800G 2xSR4/SR8 OSFP PAM4 850nm 50m DOM Two MPO-12/APC InfiniBand NDR Transceiver Module
$Quote1 Piece(MOQ)
The NVIDIA MMA4Z00-NS is a high-performance 800G InfiniBand NDR transceiver module designed for ultra-fast data center and HPC connectivity. It features dual MPO-12/APC connectors supporting 2xSR4 or SR8 PAM4 transmission over 850nm VCSEL optics, delivering up to 800Gbps bandwidth. Compatible with Broadcom VCSEL technology, it ensures low latency, high reliability, and energy efficiency. Housed in an OSFP form factor with Digital Optical Monitoring (DOM), it enables real-time diagnostics and seamless integration into next-gen InfiniBand networks. Ideal for AI, ML, and high-throughput computing environments, it offers unmatched speed, scalability, and performance.
Bulk Order Discounts Available -
NVIDIA RTX 5880 Ada Generation
$Quote1 Piece(MOQ)
The NVIDIA RTX 5880 Ada Generation is a high-performance professional graphics card designed for advanced AI, rendering, and simulation workloads. Built on the Ada Lovelace architecture, it features 48GB of GDDR6 ECC memory, 9,728 CUDA cores, and third-generation RT Cores and fourth-generation Tensor Cores for accelerated ray tracing and AI performance. With support for PCIe Gen 4 and NVIDIA NVLink, it delivers exceptional scalability and bandwidth. Ideal for data scientists, engineers, and content creators, the RTX 5880 offers unmatched computational power, energy efficiency, and real-time rendering capabilities, making it a top-tier solution for demanding professional applications.
Bulk Order Discounts Available
Bulk Order Discounts Available
