The MLA100 PCIe Card is a PCIe-based AI accelerator powered by the ARIES NPU, delivering high-performance AI inference the moment it is installed into an existing server or industrial PC. Providing 80 TOPS at just 25W — approximately 82% more power-efficient than equivalent GPU solutions — it supports over 490 AI models including a wide range of large language models (LLMs). With a PCI Express Gen4 8-lane interface and up to 32GB of LPDDR4X memory, it can be deployed immediately without any infrastructure changes. Certified under KC, FCC, CE, and RoHS international quality and safety standards, the MLA100 ensures the reliability demanded by industrial environments and is optimized for edge applications across traffic management, logistics, and smart factory deployments.

The MLX-A1 Edge AI Box is an all-in-one standalone edge AI system powered by ARIES. Delivering 80 TOPS at 70W, it comes pre-loaded with a configured OS and Mobilint's software development kit, SDK qb, enabling immediate AI operation without the need for a separate server. Even in network-unstable environments, it reliably runs computer vision, LLMs, and multimodal AI models, making it suitable for a wide range of applications from factory floors to research laboratories.

The REGULUS is a compact on-device AI system-on-chip. By integrating all essential components for AI inference — including NPU, CPU, ISP, and codec — into a single chip, it delivers a fully autonomous AI system without relying on any external processing unit. Achieving 10 TOPS of compute performance at under 3W of ultra-low power consumption, REGULUS is optimized for a wide range of on-device environments where power and space are limited, including drones, AI CCTV, compact robotics, and IoT devices.

Mobilint delivers production-ready, high-performance AI inference solutions for every layer of the edge — from server-level acceleration to the smallest embedded device