{"title":"Nvidia","description":"\u003cp\u003eMellanox is a pioneer in high-performance networking for data centers and is fully integrated into NVIDIA. The company developed Ethernet and InfiniBand technologies that enable extremely high data rates and minimal latency.\u003c\/p\u003e\n\u003cp\u003eToday, Mellanox forms the core of NVIDIA's networking division and powers AI clusters, cloud infrastructures, and supercomputers worldwide – with SmartNICs, switches, and interconnect solutions that move data between GPUs and servers at lightning speed.\u003c\/p\u003e\n\u003cp\u003eA partner for maximum performance when it comes to AI, high-performance computing, and modern data center architectures.\u003c\/p\u003e","products":[{"product_id":"mellanox-nvidia-connectx-5-en-mcx512a-acat-dual-port-10-25gbe-sfp28-netzwerkadapter-pcie-3-0-x8-low-profile","title":"Mellanox NVIDIA ConnectX-5 EN (MCX512A-ACAT) – Dual-Port 10\/25GbE SFP28 Network Adapter (PCIe 3.0 x8, Low Profile)","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-5 EN (MCX512A-ACAT)\u003c\/strong\u003e is a network adapter with two SFP28 ports for 10 or 25 Gigabit Ethernet. It is suitable for servers in data centers, virtualization and cloud environments and is connected via PCIe 3.0 x8.\u003c\/p\u003e\n\u003cp\u003eThe adapter supports RDMA over Converged Ethernet (RoCE), NVMe over Fabrics, SR-IOV, as well as offloads for VXLAN and Geneve, thereby offloading the CPU. Drivers are available for Windows Server, common Linux distributions, and VMware ESXi.\u003c\/p\u003e\n\u003cp\u003eIt can be connected using SFP28 direct attach cables (DAC) or optical SFP28\/SFP+ modules. The card is a low-profile variant; depending on the scope of delivery, a full-height bracket is included.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58594616934664,"sku":"MCX512A-ACAT","price":392.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-nvidia-connectx-5-en-mcx512a-acat-1.jpg?v=1794158167"},{"product_id":"nvidia-mellanox-connectx-4-lx-en-dual-port-25gbe-sfp28-netzwerkadapter-pcie-3-0-x8","title":"NVIDIA Mellanox ConnectX-4 Lx EN – Dual-Port 25GbE SFP28 Network Adapter, PCIe 3.0 x8","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-4 Lx EN MCX4121A-ACAT\u003c\/strong\u003e is a high-performance dual-port network adapter for servers, storage systems, virtualization platforms, and data centers. The card features two SFP28 ports, each supporting up to 25 Gigabit Ethernet, and connects via PCI Express 3.0 x8.\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eIt supports 1\/10\/25GbE connections, RDMA over Converged Ethernet (RoCE), SR-IOV, hardware offloads, and overlay networks such as VXLAN, NVGRE, and GENEVE. This makes the card particularly suitable for fast storage networks, iSCSI\/NVMe-over-Fabrics, virtual machines, cloud, and high-performance computing environments.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58595146498312,"sku":"MCX4121A-ACAT","price":152.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-mellanox-mcx4121a-acat-main.jpg?v=1794165616"},{"product_id":"nvidia-mellanox-connectx-5-vpi-100gbe-edr-infiniband-netzwerkadapter-1-qsfp28-pcie-3-0-x16-roce-rdma-mcx555a-ecat","title":"NVIDIA Mellanox ConnectX‑5 VPI – 100GbE\/EDR InfiniBand Network Adapter, 1× QSFP28, PCIe 3.0 x16, RoCE, RDMA, MCX555A‑ECAT","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-5 VPI (MCX555A-ECAT)\u003c\/strong\u003e is a server network adapter that provides either \u003cstrong\u003eInfiniBand EDR (100 Gb\/s)\u003c\/strong\u003e or \u003cstrong\u003e100-Gigabit Ethernet\u003c\/strong\u003e via a single \u003cstrong\u003eQSFP28 port\u003c\/strong\u003e (VPI = Virtual Protocol Interconnect). It connects via \u003cstrong\u003ePCIe 3.0 x16\u003c\/strong\u003e.\u003c\/p\u003e\n\u003cp\u003eThe adapter is based on the ConnectX-5 architecture and supports RDMA (InfiniBand and RoCE v1\/v2), SR-IOV, NVMe over Fabrics, as well as offloads for VXLAN, NVGRE, and Geneve. It is suitable for HPC clusters, AI systems, storage networks, and virtualized environments with high throughput and low-latency requirements.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58595885842696,"sku":"MCX555A-ECAT","price":615.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-mellanox-mcx555a-ecat-main.jpg?v=1794166317"},{"product_id":"nvidia-mellanox-connectx-5-ex-en-100gbe-netzwerkadapter-2-qsfp28-pcie-4-0-x16-roce-rdma-mcx516a-cdat","title":"NVIDIA Mellanox ConnectX‑5 Ex EN – 100GbE Network Adapter, 2× QSFP28, PCIe 4.0 x16, RoCE, RDMA, MCX516A‑CDAT","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX‑5 Ex EN (MCX516A‑CDAT)\u003c\/strong\u003e is a high-performance dual-port 100-Gigabit Ethernet network adapter for servers and workstations, based on the ConnectX‑5 Ex architecture. It features two \u003cstrong\u003eQSFP28 ports\u003c\/strong\u003e supporting 1\/10\/25\/40\/50\/100 Gigabit Ethernet and is designed for \u003cstrong\u003ePCIe 4.0 x16\u003c\/strong\u003e (backward compatible with PCIe 3.0 x16).\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe adapter is suitable for AI\/ML workloads, HPC clusters, storage networks (NVMe‑oF, Ceph, etc.), cloud infrastructures, and virtualized environments with high requirements for throughput, low latency, and CPU offloading. It offers hardware offloads for RDMA (RoCE v1\/v2), VXLAN, NVGRE, GENEVE, NVMe‑oF, and other data center protocols.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58596052205832,"sku":"MCX516A-CDAT","price":722.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-connectx-5-ex-mcx516a-cdat-1.jpg?v=1793993000"},{"product_id":"nvidia-mellanox-connectx-5-en-100gbe-netzwerkadapter-2-qsfp28-pcie-3-0-x16-roce-rdma-mcx516a-ccat","title":"NVIDIA Mellanox ConnectX‑5 EN – 100GbE Network Adapter, 2× QSFP28, PCIe 3.0 x16, RoCE, RDMA, MCX516A‑CCAT","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX‑5 EN (MCX516A‑CCAT)\u003c\/strong\u003e is a high-performance dual-port 100-Gigabit Ethernet network adapter for servers and workstations based on the ConnectX‑5 architecture. It features two \u003cstrong\u003eQSFP28 ports\u003c\/strong\u003e supporting 1\/10\/25\/40\/50\/100 Gigabit Ethernet and is designed for \u003cstrong\u003ePCIe 3.0 x16\u003c\/strong\u003e (backward compatible with PCIe 2.0 x16).\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe adapter is suitable for AI\/ML workloads, HPC clusters, storage networks (NVMe‑oF, Ceph, etc.), cloud infrastructures, and virtualized environments with high requirements for throughput, low latency, and CPU offloading. It offers hardware offloads for RDMA (RoCE v1\/v2), VXLAN, NVGRE, GENEVE, NVMe‑oF, and other datacenter protocols.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58596061511944,"sku":"MCX516A-CCAT","price":748.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-mellanox-connectx-5-en-mcx516a-ccat-1.jpg?v=1794158187"},{"product_id":"nvidia-mellanox-connectx-5-en-mcx512a-acut-dual-port-25gbe-sfp28-pcie-3-0-x8-netzwerkadapter-uefi-enabled","title":"NVIDIA Mellanox ConnectX‑5 EN MCX512A‑ACUT – Dual‑Port 25GbE SFP28 PCIe 3.0 x8 Network Adapter, UEFI‑Enabled","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-5 EN (MCX512A-ACUT)\u003c\/strong\u003e is a network adapter for servers, storage systems, virtualization, and cloud environments. It features two SFP28 ports for 1\/10\/25 Gigabit Ethernet and connects via \u003cstrong\u003ePCIe 3.0 x8\u003c\/strong\u003e.\u003c\/p\u003e\n\u003cp\u003eThe adapter is based on the ConnectX-5 architecture and supports RDMA over Converged Ethernet (RoCE v1\/v2), SR-IOV, and offloads for VXLAN, NVGRE, and Geneve, significantly reducing CPU load. It also supports NVMe over Fabrics and PTP (IEEE 1588); the boot ROM is designed for UEFI. Drivers are available for Windows Server, major Linux distributions, and VMware ESXi.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58596646584584,"sku":"MCX512A-ACUT","price":298.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-mellanox-connectx-5-en-mcx512a-acut-mcx512a-acut-1.jpg?v=1794158190"},{"product_id":"nvidia-mellanox-connectx-4-lx-en-mcx4121a-xcat-dual-port-10gbe-sfp-pcie-3-0-x8-netzwerkadapter","title":"NVIDIA Mellanox ConnectX‑4 Lx EN MCX4121A‑XCAT – Dual‑Port 10GbE SFP+ PCIe 3.0 x8 Network Adapter","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-4 Lx EN (MCX4121A-XCAT)\u003c\/strong\u003e is a network adapter for servers, storage systems, virtualization, and cloud environments. It features two SFP+ ports for 1\/10 Gigabit Ethernet and connects via \u003cstrong\u003ePCIe 3.0 x8\u003c\/strong\u003e.\u003c\/p\u003e\n\u003cp\u003eThe adapter is based on the ConnectX-4 Lx architecture and supports RDMA over Converged Ethernet (RoCE v1\/v2), SR-IOV, and offloads for VXLAN, NVGRE, and Geneve, which significantly reduce CPU load. Drivers are available for Windows Server, popular Linux distributions, and VMware ESXi.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58596679516424,"sku":"MCX4121A-XCAT","price":232.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-connectx-4-lx-mcx4121a-xcat-1.jpg?v=1793993000"},{"product_id":"mellanox-switch-ib-2-msb7890-es2f-unmanaged-36-port-edr-infiniband-switch-100-gb-s-p2c-airflow-1u","title":"Mellanox Switch‑IB 2 MSB7890‑ES2F – Unmanaged 36‑Port EDR InfiniBand‑Switch (100 Gb\/s), P2C‑Airflow, 1U","description":"\u003cp\u003eThe \u003cstrong\u003eMellanox Switch-IB 2 MSB7890 (MSB7890-ES2F)\u003c\/strong\u003e is an EDR-generation InfiniBand switch for HPC clusters, AI and ML systems, storage fabrics, and other latency-critical environments. It offers \u003cstrong\u003e36 QSFP28 ports, each providing 100 Gb\/s EDR InfiniBand\u003c\/strong\u003e, and is designed as a 1U rack-mountable switch.\u003c\/p\u003e\n\u003cp\u003eThe switch is based on Mellanox's (now NVIDIA Networking) Switch-IB 2 architecture and achieves a switching capacity of up to \u003cstrong\u003e7.2 Tbps\u003c\/strong\u003e with a port-to-port latency of approximately 90 ns. The airflow is \u003cstrong\u003eP2C (Power-to-Connector)\u003c\/strong\u003e: air flows from the power supply side to the port side. This variant has \u003cstrong\u003eno onboard management\u003c\/strong\u003e (externally managed): a subnet manager is required on the network for operation, for example on a server (OpenSM) or via NVIDIA UFM.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58597043667208,"sku":"MSB7890-ES2F","price":20899.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-msb7890-infiniband-switch-1.jpg?v=1793993000"},{"product_id":"mellanox-quantum-qm8790-mqm8790-hs2f-unmanaged-40-port-hdr-infiniband-switch-200-gb-s-p2c-airflow-1u","title":"Mellanox Quantum QM8790 MQM8790‑HS2F – Unmanaged 40‑Port HDR InfiniBand‑Switch (200 Gb\/s), P2C‑Airflow, 1U","description":"\u003cp\u003eThe \u003cstrong\u003eMellanox Quantum QM8790 (MQM8790-HS2F)\u003c\/strong\u003e is an HDR-generation InfiniBand switch for HPC clusters, AI and ML systems, storage fabrics, and other latency-critical environments. It offers \u003cstrong\u003e40 QSFP56 ports, each providing 200 Gb\/s HDR InfiniBand\u003c\/strong\u003e, and is designed as a 1U rack-mountable switch.\u003c\/p\u003e\n\u003cp\u003eThe switch is based on Mellanox's Quantum architecture (now NVIDIA Networking) and achieves a switching capacity of up to \u003cstrong\u003e16 Tb\/s\u003c\/strong\u003e with a port-to-port latency of approximately 130 ns. The airflow is \u003cstrong\u003eP2C (Power-to-Connector)\u003c\/strong\u003e: air flows from the power supply side to the port side. This variant has \u003cstrong\u003eno onboard management\u003c\/strong\u003e (externally managed): a subnet manager on the network is required for operation, for example on a server (OpenSM) or via NVIDIA UFM.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58597044715784,"sku":"MQM8790-HS2F","price":19999.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-quantum-qm8790-mqm8790-hs2f-mqm8790-hs2f-1.jpg?v=1794158193"},{"product_id":"mellanox-quantum-qm8700-mqm8700-hs2f-managed-40-port-hdr-infiniband-switch-200-gb-s-p2c-airflow-1u","title":"Mellanox Quantum QM8700 MQM8700‑HS2F – Managed 40‑Port HDR InfiniBand‑Switch (200 Gb\/s), P2C‑Airflow, 1U","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eMellanox Quantum QM8700 (MQM8700‑HS2F)\u003c\/strong\u003e is a high-performance, \u003cstrong\u003emanaged\u003c\/strong\u003e HDR-generation (High Data Rate) InfiniBand switch for HPC clusters, AI\/ML workloads, storage fabrics, and other latency-critical environments. It features \u003cstrong\u003e40 QSFP56 ports\u003c\/strong\u003e, each providing \u003cstrong\u003e200 Gb\/s HDR InfiniBand\u003c\/strong\u003e, and is designed for use as a top-of-rack switch in compact 1U enclosures.\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe switch is based on Mellanox's (now NVIDIA Networking) Quantum architecture and delivers a switching capacity of up to \u003cstrong\u003e16 Tb\/s\u003c\/strong\u003e with extremely low latency, typically \u003cstrong\u003e\u0026lt;130 ns\u003c\/strong\u003e (port-to-port). It supports a \u003cstrong\u003ePower-to-Connector (P2C)\u003c\/strong\u003e airflow direction (\"B2F\" – Back-to-Front) to ensure consistent cooling in standardized racks. As a managed variant, it is designed for complex fabrics requiring subnet management, adaptive routing features, and advanced InfiniBand functionality.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58597046485256,"sku":"MQM8700-HS2F","price":23999.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-qm8700-infiniband-switch-1.jpg?v=1793992968"},{"product_id":"nvidia-mellanox-connectx-7-mcx75310aas-neat-400-gb-s-ndr-infiniband-400gbe-single-port-osfp-smartnic-pcie-5-0-x16-rdma-gpudirect-in-network-computing","title":"NVIDIA Mellanox ConnectX-7 MCX75310AAS-NEAT – 400GbE\/NDR Single-Port OSFP SmartNIC, PCIe 5.0","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-7 MCX75310AAS-NEAT\u003c\/strong\u003e is a network card from the ConnectX-7 family for data centers, HPC clusters, AI systems, and cloud infrastructures. It features an OSFP port for 400 Gb\/s (NDR InfiniBand or 400GbE) and connects via \u003cstrong\u003ePCIe 5.0 x16\u003c\/strong\u003e. By default, it operates in InfiniBand mode, but Ethernet operation is also available.\u003c\/p\u003e\n\u003cp\u003eThe card supports RDMA, NVIDIA GPUDirect RDMA for direct communication with GPUs, as well as offloads for storage protocols such as NVMe over Fabrics and for overlay networks. This enables 400 Gb\/s performance with very low latency and low CPU overhead. This variant (AAS) offers Secure Boot; hardware encryption (IPsec, TLS, MACsec) is reserved for models with \"AAC\" in the designation.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58977246511368,"sku":"MCX75310AAS-NEAT","price":1619.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-connectx-7-mcx75310aas-neat-main.jpg?v=1794167317"},{"product_id":"nvidia-mellanox-connectx-7-mcx755106as-heat-200-gb-s-ndr200-infiniband-200gbe-dual-port-qsfp112-smartnic-pcie-5-0-x16-rdma-gpudirect-in-network-computing","title":"NVIDIA Mellanox ConnectX-7 MCX755106AS-HEAT – 200GbE\/NDR200 Dual-Port QSFP112 SmartNIC, PCIe 5.0","description":"\u003cp\u003eThe \u003cstrong\u003eNVIDIA Mellanox ConnectX-7 MCX755106AS-HEAT\u003c\/strong\u003e is a network card from the ConnectX-7 family for data centers, HPC clusters, AI systems, and cloud infrastructures. It features two QSFP112 ports, each providing 200 Gb\/s (NDR200 InfiniBand or 200GbE), and connects via \u003cstrong\u003ePCIe 5.0 x16\u003c\/strong\u003e. By default, it operates in InfiniBand mode, but Ethernet operation is also available.\u003c\/p\u003e\n\u003cp\u003eThe card supports RDMA, NVIDIA GPUDirect RDMA for direct communication with GPUs, as well as offloads for storage protocols such as NVMe over Fabrics and for overlay networks. This enables 2× 200 Gb\/s with very low latency and low CPU overhead. This variant (AAS) offers Secure Boot; hardware encryption (IPsec, TLS, MACsec) is reserved for models with \"AAC\" in the designation.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58977253294344,"sku":"MCX755106AS-HEAT","price":1919.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-connectx-7-mcx755106as-heat-1.png?v=1793993000"},{"product_id":"nvidia-rtx-pro-5000-blackwell-900-5g153-2250-000-48-gb-gddr7-ecc-workstation-gpu-pcie-5-0-x16-4-displayport-2-1-300-w-dual-slot","title":"NVIDIA RTX PRO 5000 Blackwell 900‑5G153‑2250‑000 – 48 GB GDDR7 ECC Workstation‑GPU, PCIe 5.0 x16, 4× DisplayPort 2.1, 300 W, Dual‑Slot","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA RTX PRO 5000 Blackwell (900‑5G153‑2250‑000)\u003c\/strong\u003e is a high-performance professional workstation graphics card based on the Blackwell architecture for AI development, LLM inference, generative AI workflows, high-resolution video production, complex 3D modeling, and scientific computing. It features \u003cstrong\u003e48 GB of GDDR7 memory with ECC\u003c\/strong\u003e, a memory bandwidth of \u003cstrong\u003e1,344 GB\/s\u003c\/strong\u003e, and supports \u003cstrong\u003ePCIe 5.0 x16\u003c\/strong\u003e for maximum CPU-GPU data transfer.\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eBased on the \u003cstrong\u003eNVIDIA Blackwell architecture\u003c\/strong\u003e with \u003cstrong\u003e14,080 CUDA cores\u003c\/strong\u003e, \u003cstrong\u003e440 5th-generation Tensor cores\u003c\/strong\u003e, and \u003cstrong\u003e110 4th-generation Ray Tracing cores\u003c\/strong\u003e, the card enables ultra-fast AI calculations (up to \u003cstrong\u003e2,064 AI TOPS\u003c\/strong\u003e), real-time ray tracing, and DLSS 4 Multi-Frame Generation. With \u003cstrong\u003e3× 9th-gen NVENC\u003c\/strong\u003e and \u003cstrong\u003e3× 6th-gen NVDEC\u003c\/strong\u003e, it is optimized for professional video encoding\/decoding workflows and offers \u003cstrong\u003e4× DisplayPort 2.1\u003c\/strong\u003e for up to 8K resolution at 100 Hz.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58977255588104,"sku":"900-5G153-2250-000","price":8969.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-rtx-pro-5000-blackwell-1.jpg?v=1793993026"},{"product_id":"nvidia-rtx-2000-ada-900-5g192-2240-000-16-gb-gddr6-ecc-workstation-gpu-pcie-4-0-x8-4-mini-displayport-1-4a-70-w-low-profile-dual-slot","title":"NVIDIA RTX 2000 Ada 900‑5G192‑2240‑000 – 16 GB GDDR6 ECC Workstation‑GPU, PCIe 4.0 x8, 4× Mini‑DisplayPort 1.4a, 70 W, Low‑Profile Dual‑Slot","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA RTX 2000 Ada (900‑5G192‑2240‑000)\u003c\/strong\u003e is a compact, energy-efficient professional workstation graphics card based on the Ada Lovelace architecture for CAD, DCC, BIM, medical imaging, light AI inference, and multi-monitor environments. It features \u003cstrong\u003e16 GB of GDDR6 memory with ECC\u003c\/strong\u003e, a memory bandwidth of \u003cstrong\u003e224 GB\/s\u003c\/strong\u003e, and supports \u003cstrong\u003ePCIe 4.0 x8\u003c\/strong\u003e (electrical) \/ \u003cstrong\u003ex16\u003c\/strong\u003e (mechanical) for flexible integration into space-saving workstations and servers.\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eBased on the \u003cstrong\u003eNVIDIA Ada Lovelace architecture\u003c\/strong\u003e with \u003cstrong\u003e2,816 CUDA cores\u003c\/strong\u003e, \u003cstrong\u003e88 4th-generation Tensor cores\u003c\/strong\u003e, and \u003cstrong\u003e22 3rd-generation RT cores\u003c\/strong\u003e, the card enables accelerated AI calculations (up to \u003cstrong\u003e191.9 TFLOPS\u003c\/strong\u003e with Tensor Cores), real-time ray tracing, and DLSS 3 Frame Generation. With \u003cstrong\u003e4× Mini DisplayPort 1.4a\u003c\/strong\u003e (with locking mechanism), it is optimized for driving up to four 4K or two 8K displays and offers a typical power consumption of just \u003cstrong\u003e70 W\u003c\/strong\u003e without an additional PCIe power connector.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58977322369288,"sku":"900-5G192-2240-000","price":839.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-rtx-2000-ada-16gb-1.jpg?v=1793993000"},{"product_id":"nvidia-rtx-pro-4000-blackwell-900-5g147-2270-000-24-gb-gddr7-ecc-workstation-gpu-pcie-5-0-x16-4-displayport-2-1b-145-w-single-slot","title":"NVIDIA RTX PRO 4000 Blackwell 900‑5G147‑2270‑000 – 24 GB GDDR7 ECC Workstation‑GPU, PCIe 5.0 x16, 4× DisplayPort 2.1b, 145 W, Single‑Slot","description":"\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eThe \u003cstrong\u003eNVIDIA RTX PRO 4000 Blackwell (900‑5G147‑2270‑000)\u003c\/strong\u003e is a compact, energy-efficient professional workstation graphics card based on the Blackwell architecture for AI development, LLM inference, generative AI workflows, high-resolution video production, complex 3D modeling, and scientific computing. It features \u003cstrong\u003e24 GB of GDDR7 memory with ECC\u003c\/strong\u003e, a memory bandwidth of \u003cstrong\u003e672 GB\/s\u003c\/strong\u003e, and supports \u003cstrong\u003ePCIe 5.0 x16\u003c\/strong\u003e for maximum CPU-GPU data transfer.\u003c\/p\u003e\n\u003cp class=\"my-2 [\u0026amp;+p]:mt-4 [\u0026amp;_strong:has(+br)]:inline-block [\u0026amp;_strong:has(+br)]:align-top\"\u003eBased on the \u003cstrong\u003eNVIDIA Blackwell architecture\u003c\/strong\u003e with \u003cstrong\u003e8,960 CUDA cores\u003c\/strong\u003e, \u003cstrong\u003e280 5th-generation Tensor cores\u003c\/strong\u003e, and \u003cstrong\u003e70 4th-generation Ray Tracing cores\u003c\/strong\u003e, the card enables ultra-fast AI computations (up to \u003cstrong\u003e770 AI TOPS\u003c\/strong\u003e), real-time ray tracing, and DLSS 4 Multi-Frame Generation. With \u003cstrong\u003e2× 9th-gen NVENC\u003c\/strong\u003e and \u003cstrong\u003e2× 6th-gen NVDEC\u003c\/strong\u003e, it is optimized for professional video encoding\/decoding workflows and offers \u003cstrong\u003e4× DisplayPort 2.1b\u003c\/strong\u003e for up to 8K resolution at 240 Hz.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":58977361232136,"sku":"900-5G147-2270-000","price":2949.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/nvidia-rtx-pro-4000-blackwell-1.png?v=1793993000"},{"product_id":"mellanox-nvidia-connectx-6-lx-mcx631102an-adat-dual-port-25gbe-sfp28-netzwerkadapter-pcie-4-0-x8-low-profile","title":"Mellanox NVIDIA ConnectX-6 Lx (MCX631102AN-ADAT) – Dual-Port 25GbE SFP28 Netzwerkadapter (PCIe 4.0 x8, Low Profile)","description":"\u003cp\u003eDer \u003cstrong\u003eNVIDIA Mellanox ConnectX-6 Lx (MCX631102AN-ADAT)\u003c\/strong\u003e ist ein Netzwerkadapter mit \u003cstrong\u003ezwei SFP28-Ports\u003c\/strong\u003e für \u003cstrong\u003e1\/10\/25 Gigabit-Ethernet\u003c\/strong\u003e und wird über \u003cstrong\u003ePCIe 4.0 x8\u003c\/strong\u003e angebunden. Die Low-Profile-Karte (HHHL) wird mit montierter Full-Height-Blende geliefert, eine Low-Profile-Blende liegt bei.\u003c\/p\u003e\n\u003cp\u003eMit \u003cstrong\u003eRoCE\u003c\/strong\u003e, \u003cstrong\u003eSR-IOV\u003c\/strong\u003e, ASAP²-Offload für Open vSwitch, NVMe-oF-Target-Offload sowie VXLAN-\/NVGRE-Offloads entlastet er die CPU in Virtualisierungs-, Cloud- und Storage-Umgebungen. Angeschlossen wird er über SFP28-Direktkabel (DAC) oder optische SFP28\/SFP+-Module.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":59421524066568,"sku":"MCX631102AN-ADAT","price":411.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-nvidia-connectx-6-lx-mcx631102an-adat-1.jpg?v=1794235485"},{"product_id":"mellanox-nvidia-connectx-6-lx-mcx631102as-adat-dual-port-25gbe-sfp28-netzwerkadapter-pcie-4-0-x8-low-profile-secure-boot","title":"Mellanox NVIDIA ConnectX-6 Lx (MCX631102AS-ADAT) – Dual-Port 25GbE SFP28 Netzwerkadapter (PCIe 4.0 x8, Low Profile, Secure Boot)","description":"\u003cp\u003eDer \u003cstrong\u003eNVIDIA Mellanox ConnectX-6 Lx (MCX631102AS-ADAT)\u003c\/strong\u003e ist ein Netzwerkadapter mit \u003cstrong\u003ezwei SFP28-Ports\u003c\/strong\u003e für \u003cstrong\u003e1\/10\/25 Gigabit-Ethernet\u003c\/strong\u003e und wird über \u003cstrong\u003ePCIe 4.0 x8\u003c\/strong\u003e angebunden. Die Low-Profile-Karte (HHHL) wird mit montierter Full-Height-Blende geliefert, eine Low-Profile-Blende liegt bei.\u003c\/p\u003e\n\u003cp\u003eMit \u003cstrong\u003eRoCE\u003c\/strong\u003e, \u003cstrong\u003eSR-IOV\u003c\/strong\u003e, ASAP²-Offload für Open vSwitch, NVMe-oF-Target-Offload sowie VXLAN-\/NVGRE-Offloads entlastet er die CPU in Virtualisierungs-, Cloud- und Storage-Umgebungen. Über den hardwarebasierten \u003cstrong\u003eSecure Boot\u003c\/strong\u003e (Hardware Root of Trust) wird nur signierte Firmware ausgeführt. Angeschlossen wird er über SFP28-Direktkabel (DAC) oder optische SFP28\/SFP+-Module.\u003c\/p\u003e","brand":"Mellanox","offers":[{"title":"Default Title","offer_id":59421526196488,"sku":"MCX631102AS-ADAT","price":450.99,"currency_code":"EUR","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/files\/mellanox-nvidia-connectx-6-lx-mcx631102as-adat-1.jpg?v=1794235496"}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0987\/9217\/7928\/collections\/b872419e.png?v=1793993750","url":"https:\/\/luniumsystems.de\/en\/collections\/nvidia-mellanox.oembed","provider":"Lunium","version":"1.0","type":"link"}