{"title":"NVIDIA at Knoxlabs","description":"\u003cp\u003eExplore NVIDIA Jetson developer kits and reference compute options. Compare the platform, connectivity and software workflow for perception, robotics and AI development.\u003c\/p\u003e","products":[{"product_id":"nvidia-dgx-spark","title":"NVIDIA DGX Spark (Reference Only)","description":"\u003cp\u003e\u003cstrong\u003eReference only.\u003c\/strong\u003e Knoxlabs does not sell or ship the NVIDIA DGX Spark. This page is provided for product information.\u003c\/p\u003e\u003cp\u003eA deskside training box with 128 GB of coherent unified memory. In a robot-learning programme this is the machine that sits beside the capture rig: you post-train or fine-tune a policy on the demonstrations you just collected, then push the result to the Jetson on the robot.\u003c\/p\u003e\u003ch3\u003eSpecifications\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003eUp to 1 petaFLOP of FP4 AI performance\u003c\/li\u003e\n\u003cli\u003eNVIDIA GB10 Grace Blackwell superchip\u003c\/li\u003e\n\u003cli\u003e20-core Arm CPU (10× Cortex-X925, 10× Cortex-A725)\u003c\/li\u003e\n\u003cli\u003e128 GB coherent unified LPDDR5x at 273 GB\/s\u003c\/li\u003e\n\u003cli\u003e4 TB self-encrypting NVMe storage\u003c\/li\u003e\n\u003cli\u003eConnectX-7 200 Gbps networking for two-unit clustering\u003c\/li\u003e\n\u003cli\u003e240 W power supply, 140 W chip TDP\u003c\/li\u003e\n\u003cli\u003eIncludes a 90-day NVIDIA AI Enterprise licence\u003c\/li\u003e\n\u003c\/ul\u003e\u003ch3\u003eWhere it fits\u003c\/h3\u003e\u003cp\u003eLayer 7 of the stack — see \u003ca href=\"\/pages\/robotics-data\"\u003edata \u0026amp; policy\u003c\/a\u003e for how collected episodes become a trained policy.\u003c\/p\u003e\u003cp\u003eSee also \u003ca href=\"\/pages\/robotics-stations\"\u003eteleoperation stations\u003c\/a\u003e and the \u003ca href=\"\/pages\/robotics\"\u003erobotics hub\u003c\/a\u003e.\u003c\/p\u003e","brand":"NVIDIA","offers":[{"title":"Default Title","offer_id":53021876453696,"sku":null,"price":5873.75,"currency_code":"USD","in_stock":false}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0704\/2759\/files\/nvidia-dgx-spark-featured-white-square_570239d6-22ee-41bd-98d7-33c57d0eca99.webp?v=1786689758"},{"product_id":"nvidia-jetson-agx-orin-developer-kit-64gb","title":"NVIDIA Jetson AGX Orin Developer Kit (64GB)","description":"\u003cp\u003eOn-robot compute for a full-size rig. 275 TOPS and 64 GB of LPDDR5 is enough to run multi-camera perception, an Isaac ROS graph and policy inference on the robot itself, instead of tethering it to a workstation.\u003c\/p\u003e\u003ch3\u003eSpecifications\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e275 sparse INT8 TOPS\u003c\/li\u003e\n\u003cli\u003e2048-core NVIDIA Ampere GPU with 64 Tensor Cores\u003c\/li\u003e\n\u003cli\u003e12-core Arm Cortex-A78AE CPU up to 2.2 GHz\u003c\/li\u003e\n\u003cli\u003e64 GB 256-bit LPDDR5 at 204.8 GB\/s\u003c\/li\u003e\n\u003cli\u003e64 GB eMMC 5.1 plus microSD expansion\u003c\/li\u003e\n\u003cli\u003eConfigurable 15 W-60 W power envelope\u003c\/li\u003e\n\u003cli\u003e110 × 110 × 71.65 mm\u003c\/li\u003e\n\u003cli\u003eRuns NVIDIA JetPack and Isaac ROS\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eSee also \u003ca href=\"\/pages\/robotics-data\"\u003edata \u0026amp; policy\u003c\/a\u003e, \u003ca href=\"\/pages\/robotics-cameras\"\u003eperception \u0026amp; cameras\u003c\/a\u003e and the \u003ca href=\"\/pages\/robotics\"\u003erobotics hub\u003c\/a\u003e.\u003c\/p\u003e","brand":"NVIDIA","offers":[{"title":"Default Title","offer_id":53021876486464,"sku":null,"price":0.1,"currency_code":"USD","in_stock":false}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0704\/2759\/files\/nvidia-jetson-agx-orin-developer-kit-64gb-featured-white-square_61b6662e-f2cd-443a-8edd-68260f4257ff.webp?v=1786689763"},{"product_id":"nvidia-jetson-orin-nano-super-developer-kit","title":"NVIDIA Jetson Orin Nano Super Developer Kit","description":"\u003cp\u003eThe entry point into on-robot inference. 67 TOPS in a 100 × 79 mm board at 7-25 W — enough for stereo perception and a small policy on a teleoperation arm, a data-collection cart or a compact mobile base.\u003c\/p\u003e\u003ch3\u003eSpecifications\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e67 sparse INT8 TOPS (33 dense TOPS)\u003c\/li\u003e\n\u003cli\u003e512-core NVIDIA Ampere GPU with 16 Tensor Cores\u003c\/li\u003e\n\u003cli\u003e6-core Arm Cortex-A78AE CPU up to 1.7 GHz\u003c\/li\u003e\n\u003cli\u003e8 GB 128-bit LPDDR5 at 102 GB\/s\u003c\/li\u003e\n\u003cli\u003e7 W, 15 W and 25 W power modes\u003c\/li\u003e\n\u003cli\u003eM.2 Key M NVMe slot plus SD card\u003c\/li\u003e\n\u003cli\u003e100 × 79 × 21 mm\u003c\/li\u003e\n\u003cli\u003eSupports NVIDIA JetPack and Isaac ROS\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eSee also \u003ca href=\"\/pages\/robotics-data\"\u003edata \u0026amp; policy\u003c\/a\u003e, \u003ca href=\"\/pages\/robotics-cameras\"\u003eperception \u0026amp; cameras\u003c\/a\u003e and the \u003ca href=\"\/pages\/robotics\"\u003erobotics hub\u003c\/a\u003e.\u003c\/p\u003e","brand":"NVIDIA","offers":[{"title":"Default Title","offer_id":53021876519232,"sku":null,"price":0.1,"currency_code":"USD","in_stock":false}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0704\/2759\/files\/nvidia-jetson-orin-nano-super-developer-kit-featured-white-square_350d2834-190a-4606-921d-1792244c790d.webp?v=1786689769"},{"product_id":"nvidia-jetson-agx-thor-developer-kit","title":"NVIDIA Jetson AGX Thor Developer Kit","description":"\u003cp\u003eThe flagship on-robot compute for humanoid and multi-arm platforms. 2,070 FP4 TFLOPS and 128 GB of LPDDR5X means a large vision-language-action model and many synchronised camera streams can run in real time on the robot rather than off it.\u003c\/p\u003e\u003ch3\u003eSpecifications\u003c\/h3\u003e\u003cul\u003e\n\u003cli\u003e2,070 FP4 TFLOPS sparse AI performance\u003c\/li\u003e\n\u003cli\u003e2560-core NVIDIA Blackwell GPU with 10 TPCs\u003c\/li\u003e\n\u003cli\u003e14-core Arm Neoverse-V3AE at 2.6 GHz\u003c\/li\u003e\n\u003cli\u003e128 GB LPDDR5X at 273 GB\/s\u003c\/li\u003e\n\u003cli\u003e1 TB NVMe M.2 onboard storage\u003c\/li\u003e\n\u003cli\u003e40 W-130 W configurable power envelope\u003c\/li\u003e\n\u003cli\u003eRuns NVIDIA Isaac and Holoscan software stacks\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eAlso available as the T5000 production system-on-module for integration into a custom carrier board — ask us about volume terms.\u003c\/p\u003e\u003cp\u003eSee also \u003ca href=\"\/pages\/robotics-data\"\u003edata \u0026amp; policy\u003c\/a\u003e, \u003ca href=\"\/pages\/robotics-arms\"\u003earms \u0026amp; platforms\u003c\/a\u003e and the \u003ca href=\"\/pages\/robotics\"\u003erobotics hub\u003c\/a\u003e.\u003c\/p\u003e","brand":"NVIDIA","offers":[{"title":"Default Title","offer_id":53021876552000,"sku":null,"price":0.1,"currency_code":"USD","in_stock":false}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0704\/2759\/files\/nvidia-jetson-agx-thor-developer-kit-featured-white-square_2f906232-7428-4ab6-9c29-5417ecc6ce82.webp?v=1786689767"}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0704\/2759\/collections\/nvidia-jetson-agx-thor-developer-kit-featured-white-square_2f906232-7428-4ab6-9c29-5417ecc6ce82.webp?v=1791564454","url":"https:\/\/www.knoxlabs.com\/collections\/nvidia.oembed","provider":"Knoxlabs","version":"1.0","type":"link"}