Jetson Xavier NX delivers up to 21 TOPS, making it ideal for high-performance compute and AI in embedded and edge systems. You get the performance of 384 NVIDIA CUDA^® Cores, 48 Tensor Cores, 6 Carmel ARM CPUs, and two NVIDIA Deep Learning Accelerators (NVDLA) engines. Combined with over 59.7GB/s of memory bandwidth, video encoded, and decode, these features make Jetson Xavier NX the platform of choice to run multiple modern neural networks in parallel and process high-resolution data from multiple sensors simultaneously.

• Jetson Xavier NX supports multiple power modes, including low-power modes for battery-operated systems, and delivers up to 14 TOPs for AI applications in as little as 10 W
• New NVIDIA VoltaTM architecture with 384 NVIDIA CUDA cores and 48 Tensor cores, 6-core NVIDIA Carmel ARM^®v8.2 64-bit CPU 6MB L2 + 4MB L3
• Enable multiple concurrent AI applications: on-board 8 GB 128-bit LPDDR4x 59.7GB/s
• Vision Accelerator: Up to 2x Programmable Vision Accelerator
• CSI Camera: Up to 6 cameras (24 via virtual channels), 14 lanes (3x4 or 6x2 or 3x4 + 1x2 or 5x2 + 1x4) MIPI CSI-2
  D-PHY 1.2 (up to 30 Gbps)
• Include next-generation vision accelerator engine: PVA v2, and multi-standard video encoder & decoder