{"title":"Módulos embebidos y placas de desarrollo","description":"\u003cp\u003eRobots, cámaras y máquinas de diseño propio llevan dentro un módulo NVIDIA Jetson o una placa FPGA, y es el ingeniero quien los integra. Para robótica tenemos el módulo Jetson Thor con GPU Blackwell. El Jetson AGX Orin 64GB Developer Kit, con hasta 275 TOPS, está pensado para proyectos de visión artificial, mientras que la placa FPGA Kintex UltraScale sirve para prototipar lógica de alta velocidad y procesamiento de señal.\u003c\/p\u003e","products":[{"product_id":"nvidia-jetson-thor-ai-robotics-module-blackwell-gpu-128-gb","title":"NVIDIA Jetson Thor AI Robotics Module — Blackwell GPU, 128 GB","description":"\u003cp\u003eThe Jetson Thor is a system-on-module (SoM) built for physical AI and robotics at the edge, pairing an NVIDIA Blackwell architecture GPU with a 14-core Arm Neoverse-V3AE CPU. It runs the latest generative AI, vision language action (VLA), large language and vision language models locally, so robots, autonomous machines and visual-agent systems make decisions without a round trip to the cloud. Power is configurable between 40 W and 130 W, letting integrators tune the same module for a battery-powered mobile robot or a rack-fed inference node.\u003c\/p\u003e\u003cp\u003eFor teams building on top of it, Thor runs the full NVIDIA AI software stack — Isaac for robotics, Metropolis for visual agentic AI, and Holoscan for sensor processing — alongside a hardware accelerator suite that keeps real-time workloads moving:\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003eBlackwell GPU with fifth-gen Tensor Cores, a Transformer engine, and Multi-Instance GPU (MIG) support for running several models side by side.\u003c\/li\u003e\n\u003cli\u003eThird-generation Programmable Vision Accelerator (PVA), dual video encoders and decoders, and an optical flow accelerator for sensor fusion.\u003c\/li\u003e\n\u003cli\u003e4x 25 GbE networking plus extensive camera, PCIe Gen5 and audio I\/O for multi-sensor rigs — up to 20 cameras via HSB and up to 32 using virtual channels.\u003c\/li\u003e\n\u003cli\u003e128 GB of 256-bit LPDDR5X memory at 273 GB\/s on the top module, with NVMe storage over PCIe.\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eIt is offered in two modules: the T5000, which reaches up to 2070 FP4 TFLOPS with a 2560-core GPU and 128 GB of memory, and the T4000, with a 1536-core GPU, 64 GB of memory and a 40–70 W envelope. Both ship as a 100 x 87 mm module with a 699-pin board-to-board connector and an integrated Thermal Transfer Plate. Linux is the supported operating system.\u003c\/p\u003e","brand":"BarEquip.Pro","offers":[{"title":"33927 Jetson Thor T4000","offer_id":67963149090915,"sku":"BEP-0885-02","price":4500.0,"currency_code":"USD","in_stock":true},{"title":"32815 Jetson Thor T5000","offer_id":67963149123683,"sku":"BEP-0885-01","price":7300.0,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0643\/6726\/3843\/files\/7b922f8d-b1e9-4d6f-a5d9-e0cf3ab1bd75.jpg?v=1795092926"},{"product_id":"nvidia-jetson-agx-orin-64gb-ai-embedded-module-developer-kit","title":"NVIDIA Jetson AGX Orin 64GB AI Embedded Module \u0026 Developer Kit","description":"\u003cp\u003eThe Jetson AGX Orin 64GB is an edge AI compute module built for running neural-network inference where the data is created, without a round-trip to a cloud server. It pairs a 2048-core NVIDIA Ampere GPU (64 Tensor Cores) with a 12-core Arm Cortex-A78AE CPU and 64GB of 256-bit LPDDR5 memory, reaching up to 275 TOPS of AI performance. For a bar, restaurant or retail operator, that is enough headroom to run people-counting, queue analytics, loss-prevention vision or a self-order kiosk model locally on one board.\u003c\/p\u003e\u003cp\u003eIt ships as a developer kit or as the bare module for integration into a finished product, programmed through the JetPack SDK. Two NVDLA v2 deep-learning accelerators and a PVA vision accelerator offload work from the GPU, and the 16-lane MIPI CSI-2 interface drives up to six cameras for multi-feed analytics.\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003eUp to 275 TOPS AI performance, 2048-core Ampere GPU with 64 Tensor Cores\u003c\/li\u003e\n\u003cli\u003e12-core Arm Cortex-A78AE CPU at up to 2.2 GHz\u003c\/li\u003e\n\u003cli\u003e64GB 256-bit LPDDR5 (204.8 GB\/s), 64GB eMMC 5.1 storage\u003c\/li\u003e\n\u003cli\u003eUp to 10 GbE networking, PCIe Gen4, and DisplayPort 1.4a output\u003c\/li\u003e\n\u003cli\u003eConfigurable 15–75 W power envelope, programmed via the JetPack SDK\u003c\/li\u003e\n\u003c\/ul\u003e","brand":"BarEquip.Pro","offers":[{"title":"Default Title","offer_id":67963252998243,"sku":"BEP-0892","price":2375.0,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0643\/6726\/3843\/files\/f8f11b83-0add-4593-a60d-bf57af0c5b66.jpg?v=1795095761"},{"product_id":"kintex-ultrascale-fpga-development-board-xcku040-4-gb-ddr4","title":"Kintex UltraScale FPGA Development Board — XCKU040, 4 GB DDR4","description":"\u003cp\u003eThe AXKU041 is a development board built around the AMD Kintex UltraScale XCKU040-2FFVA1156 FPGA, aimed at engineering teams prototyping high-speed logic, signal processing and data-acquisition designs. It pairs the FPGA with four 1 GB DDR4 chips (64-bit bus, 2400 Mbps) and 128 Mbit QSPI flash for user data, so a design has both working memory and non-volatile storage on board from day one.\u003c\/p\u003e\u003cp\u003eConnectivity is where this board earns its place on the bench. It carries three FMC expansion sites and a mix of serial and network interfaces, which lets a lab add daughter cards for ADC\/DAC, camera or custom I\/O without redesigning the carrier.\u003c\/p\u003e\u003cul\u003e\n\u003cli\u003eFPGA: AMD Kintex UltraScale XCKU040-2FFVA1156\u003c\/li\u003e\n\u003cli\u003eMemory: four 1 GB DDR4 (64-bit bus, 2400 Mbps) plus 128 Mbit QSPI flash\u003c\/li\u003e\n\u003cli\u003eExpansion: 2× FMC LPC and 1× FMC HPC\u003c\/li\u003e\n\u003cli\u003eHigh-speed I\/O: SFP+ up to 12.5 Gb\/s, PCIe, 10\/100\/1000M Ethernet (RJ-45)\u003c\/li\u003e\n\u003cli\u003eDebug: 10-pin 0.1-inch standard JTAG port; on-board LM75 temperature sensor\u003c\/li\u003e\n\u003c\/ul\u003e\u003cp\u003eIt suits R\u0026amp;D labs, universities and hardware startups that need a ready carrier for UltraScale development rather than spinning their own board.\u003c\/p\u003e","brand":"BarEquip.Pro","offers":[{"title":"Default Title","offer_id":67963271643235,"sku":"BEP-0894","price":1686.0,"currency_code":"USD","in_stock":true}],"thumbnail_url":"\/\/cdn.shopify.com\/s\/files\/1\/0643\/6726\/3843\/files\/f84c172e-f7b2-48bf-80d5-2fd56c4ea179.jpg?v=1795096669"}],"url":"https:\/\/barequip.pro\/es\/collections\/dev-boards.oembed","provider":"BarEquip.Pro","version":"1.0","type":"link"}