Mobilint MLX-A1 Edge AI Box — 80 TOPS Industrial AI Mini PC

Specifications

Overview

The MLX-A1 is a standalone AI mini PC that pairs an Intel Core i5-13600HE host processor with Mobilint's MLA100 MXM module, adding 80 TOPS of ARIES NPU inference on top of the x86 platform. It is engineered to run large language models, vision pipelines and multimodal applications entirely offline, so data never leaves the deployment site.

The system is specified around a 70 W thermal design power for the accelerator with a 19 V DC supply rated to 230 W maximum, and is rated for -20 C to +60 C operation with 0.5 m/s of airflow. Three 2.5 GbE ports, HDMI 2.1 output at up to 4096 x 2304 @ 60 Hz, 32 GB of DDR5 and a 512 GB NVMe boot drive make it a deployable edge node rather than a development board.

Mobilint positions the box as bridging industrial-grade robustness and desktop accessibility: it arrives with Ubuntu 22.04 and the SDK qb driver and runtime preinstalled, so compiled models from the ARIES toolchain run without re-integration.

QS Compute supplies the MLX-A1 alongside the wider Mobilint ARIES and REGULUS families, including the MLA100 PCIe card, MLA400 quad-NPU card, MLA100 MXM module and the REGULUS-based edge boxes.

Key Benefits

80 TOPS of NPU inference in a 1.3 kg chassis. Fully offline operation for privacy-sensitive and disconnected sites. -20 to +60 C industrial rating with airflow. Three 2.5 GbE ports for camera and sensor aggregation. Ubuntu 22.04 + SDK preinstalled.

Applications

Offline LLM and multimodal assistants at the edge; multi-camera vision analytics; industrial quality inspection; retail and hospitality AI terminals; medical imaging workstations in closed networks; defence and public-sector sites requiring air-gapped inference.

Related Products