{"product_id":"nvidia-rtx-pro-6000-blackwell-server-edition-96gb","title":"NVIDIA RTX PRO 6000 Blackwell Server Edition (96GB)","description":"\u003cp\u003eThe NVIDIA RTX PRO 6000 Blackwell Server Edition is a 96GB GDDR7 ECC graphics card for server deployments running generative AI, scientific computing, rendering and video workloads. It combines large ECC-protected graphics memory with dedicated AI and ray-tracing hardware. It suits teams that already have a compatible server chassis and need a lot of GPU memory in a single card.\u003c\/p\u003e\u003ch2\u003eWho it's for\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003eRunning large models, datasets and complex scenes that need 96GB of GPU memory\u003c\/li\u003e\n\u003cli\u003eFP4 inference on fifth-generation Tensor cores\u003c\/li\u003e\n\u003cli\u003eSharing one GPU between up to four isolated workloads with MIG\u003c\/li\u003e\n\u003cli\u003eDemanding video pipelines that use four encode and four decode engines\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eKey features\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003e\n\u003cstrong\u003e96GB GDDR7 ECC:\u003c\/strong\u003e Capacity for large models, datasets and complex scenes, with 1,597 GB\/s memory bandwidth.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eBlackwell acceleration:\u003c\/strong\u003e Fifth-generation Tensor cores support FP4 inference, while fourth-generation RT cores accelerate ray tracing.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eShared GPU resources:\u003c\/strong\u003e MIG partitions one GPU into up to four isolated instances with dedicated resources.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eProfessional media processing:\u003c\/strong\u003e Four encode and four decode engines support demanding video pipelines.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003eBuilt-in security:\u003c\/strong\u003e Confidential computing and secure boot with root of trust.\u003c\/li\u003e\n\u003cli\u003e\n\u003cstrong\u003ePassive cooling:\u003c\/strong\u003e The heatsink relies on airflow provided by the server.\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch2\u003eSpecifications\u003c\/h2\u003e\u003ctable\u003e\n\u003cthead\u003e\u003ctr\u003e\n\u003cth\u003eSpec\u003c\/th\u003e\n\u003cth\u003eDetail\u003c\/th\u003e\n\u003c\/tr\u003e\u003c\/thead\u003e\n\u003ctbody\u003e\n\u003ctr\u003e\n\u003ctd\u003eArchitecture\u003c\/td\u003e\n\u003ctd\u003eNVIDIA Blackwell\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eGPU memory\u003c\/td\u003e\n\u003ctd\u003e96 GB GDDR7 with ECC\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory interface\u003c\/td\u003e\n\u003ctd\u003e512-bit\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMemory bandwidth\u003c\/td\u003e\n\u003ctd\u003e1,597 GB\/s\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCUDA cores\u003c\/td\u003e\n\u003ctd\u003e24,064\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eTensor cores\u003c\/td\u003e\n\u003ctd\u003e752, fifth generation\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eRT cores\u003c\/td\u003e\n\u003ctd\u003e188, fourth generation\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eFP32 performance\u003c\/td\u003e\n\u003ctd\u003e120 TFLOPS\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePeak FP4 Tensor performance\u003c\/td\u003e\n\u003ctd\u003e4 PFLOPS\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePeak FP8 Tensor performance\u003c\/td\u003e\n\u003ctd\u003e2 PFLOPS\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePeak FP16 \/ BF16 Tensor performance\u003c\/td\u003e\n\u003ctd\u003e1 PFLOP\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePeak TF32 Tensor performance\u003c\/td\u003e\n\u003ctd\u003e234 TFLOPS\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePeak RT core performance\u003c\/td\u003e\n\u003ctd\u003e355 TFLOPS\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eHost interface\u003c\/td\u003e\n\u003ctd\u003ePCI Express 5.0 x16\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMulti-Instance GPU\u003c\/td\u003e\n\u003ctd\u003eUp to 4 isolated instances; up to 24 GB per instance\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eVideo engines\u003c\/td\u003e\n\u003ctd\u003e4 ninth-generation NVENC; 4 sixth-generation NVDEC\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCooling\u003c\/td\u003e\n\u003ctd\u003ePassive heatsink; requires server-provided airflow\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eMaximum board power\u003c\/td\u003e\n\u003ctd\u003eUp to 600 W, configurable\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003ePower connector\u003c\/td\u003e\n\u003ctd\u003e1 × PCIe CEM5 16-pin\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eForm factor\u003c\/td\u003e\n\u003ctd\u003eFull-height, full-length, dual-slot\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eCard dimensions\u003c\/td\u003e\n\u003ctd\u003e4.4 inches high × 10.5 inches long\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eDisplay connectors\u003c\/td\u003e\n\u003ctd\u003e4 × DisplayPort 2.1\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eNVLink\u003c\/td\u003e\n\u003ctd\u003eNot supported\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003ctr\u003e\n\u003ctd\u003eSecurity\u003c\/td\u003e\n\u003ctd\u003eConfidential computing; secure boot with root of trust\u003c\/td\u003e\n\u003c\/tr\u003e\n\u003c\/tbody\u003e\n\u003c\/table\u003e\u003ch2\u003eGood to know\u003c\/h2\u003e\u003cul\u003e\n\u003cli\u003ePerformance figures are manufacturer-rated peaks; application throughput depends on precision, software and system configuration.\u003c\/li\u003e\n\u003cli\u003eThis listing is for the passive, air-cooled Server Edition GPU card. Its heatsink depends on forced airflow from a compatible server chassis.\u003c\/li\u003e\n\u003cli\u003eConfirm chassis clearance, PCIe slot support, power cabling and cooling capacity before installation.\u003c\/li\u003e\n\u003cli\u003eIt is not a complete server.\u003c\/li\u003e\n\u003cli\u003eOptional enterprise software licensing is separate unless expressly included in your order.\u003c\/li\u003e\n\u003cli\u003eManufacturer documentation referenced by the original listing: NVIDIA product specifications, PNY specifications, and a product brief with integration requirements.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eChoose this card if your server can feed a passive, dual-slot card with forced airflow and you want 96GB of ECC memory per GPU. If you need a complete machine rather than a component, look at the full systems in the ComputaHardware range instead.\u003c\/p\u003e","brand":"NVIDIA","offers":[{"title":"Default Title","offer_id":58232536301950,"sku":"RTX-pro-6000-blackwell-server-edition-96gb","price":15339.0,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/1022\/4468\/0062\/files\/nvidia-rtx-pro-6000-blackwell-server-edition_17d67567-32ee-4b89-b48b-60403da72568.png?v=1792094006","url":"https:\/\/computahardware.net\/products\/nvidia-rtx-pro-6000-blackwell-server-edition-96gb","provider":"ComputaHardware","version":"1.0","type":"link"}