The UP Xtreme ARL AI Development Kit integrates Intel® Core™ Ultra 5 processors (Arrow Lake series) and Intel Arc graphics in a 120.35 × 122.5 mm industrial-grade form factor, designed for AI inferencing and edge computing. It supports up to 64GB dual-channel LPDDR5 (6400 MT/s) memory and dual M.2 2280 M-key PCIe Gen4 x4 slots for NVMe storage. With 83 Platform TOPS NPU performance, it enables real-time deep learning across computer vision and robotics workloads. 

Connectivity includes 2.5GbE and GbE LAN, USB 3.2 Gen2, HDMI 2.1, DP 2.1, and a 40-pin GPIO header supporting UART, I²C, I²S, ADC, and PWM for embedded I/O expansion. The board operates on 9–36V DC input, supports Win 11/10 LTSC and Ubuntu 24.04 LTS, and delivers 3× AI performance over previous generations. Optimized for scalable, high-throughput industrial and IoT applications, it provides robust AI acceleration in compact embedded environments

UP Xtreme ARL AI Development Kit Features  

Processor & Graphics Architecture

The board integrates the Intel Core Ultra 5 Processors (Arrow Lake) alongside an integrated Intel Arc Graphics engine. It offers up to 14 cores and supports dual-channel LPDDR5 memory up to 64 GB. The graphics subsystem employs XMX systolic arrays to accelerate AI workloads, contributing to enhanced rendering and inferencing performance.

AI Performance & NPU Integration

A dedicated Neural Processing Unit (NPU) is embedded to handle AI inference tasks, providing up to 83 platform TOPS of performance. The architecture offloads deep learning and computer-vision workloads from the main CPU, enabling real-time processing in embedded and edge scenarios. This makes the board suitable for robotics, automation, and vision-based applications where low latency and efficient compute matter.

Expansion & I/O Flexibility

The board supports a 40-pin HAT connector plus multiple wafer-style headers, enabling interfaces including GPIO, UART, I²C, I²S, PWM, and ADC. For storage and AI module expansion, it offers dual M.2 2280 M-key slots (PCIe Gen 4 ×4 NVMe), a M.2 2230 E-key slot for WiFi/BT and a M.2 3052 B-key slot with nano-SIM for 5G modules. Together these options allow integration of sensors, actuators, cameras and network modules for embedded system development.

Power, Display & Connectivity

Operating over a wide 9–36 V DC input range, the board is fitted with an onboard TPM 2.0 and RTC battery for industrial use cases requiring secure, always-on operation. Display output support includes up to four concurrent screens via HDMI 2.1 (×2), DP 2.1 (×1) and DP 1.4 via USB-C (×1). Network connectivity includes one 2.5 GbE and one GbE interface, with WiFi/BT and LTE/5G support via expansion.

Thermal Design & Power Efficiency

The UP Xtreme ARL integrates an active cooling solution designed to maintain stable thermal performance during high AI workloads. Typical power consumption ranges from 57.6 W to 86.4 W, with support for AT and ATX power modes. Its wide 9–36 V DC input ensures compatibility with various industrial power systems, improving reliability in rugged environments.

Environmental & Mechanical Specifications

Engineered for industrial environments, the board operates within 0°C to 60°C and tolerates 0–90% relative humidity (non-condensing). The compact 120.35 × 122.5 mm form factor supports fan and passive cooling options, depending on enclosure requirements.

Software Support & Security

It supports Windows 11 LTSC/10 LTSC and Linux (Ubuntu 24.04 LTS and Yocto 5+). Integrated TPM 2.0 hardware ensures secure boot and encryption, essential for IoT and industrial edge deployments.

Applications

The UP Xtreme ARL AI Development Kit is designed for demanding edge AI and embedded computing applications requiring high-performance inferencing and low latency. It is widely used in industrial automation, where it enables machine vision, predictive maintenance, and real-time quality inspection. In robotics, the integrated NPU and Intel® Arc™ graphics support object detection, SLAM, and motion control. The board also powers smart transportation systems, retail analytics, and medical imaging through its multi-display, sensor, and I/O flexibility. Its support for 5G, Ethernet, and AI frameworks like OpenVINO™ makes it ideal for distributed edge nodes and IoT gateways operating under mission-critical workloads.