HAILO-10H M.2 2280

The Hailo-10H is the industry's first edge AI accelerator purpose-built for on-device generative AI, delivering up to 40 TOPS INT4 performance with exceptional power efficiency at just 2.5W typical consumption. Designed for advanced AI workloads, it supports 4GB or 8GB LPDDR4/LPDDR4X memory through a dedicated DDR interface, enabling seamless deployment of Large Language Models (LLMs), Vision-Language Models (VLMs), computer vision, and multimodal AI applications directly at the edge.

Integrated with the UP ecosystem, the Hailo-10H accelerates AI inference on UP single board computers and embedded systems while maintaining complete data privacy through offline processing. Developers can run models such as Llama 2, Stable Diffusion, and real-time deep learning applications without cloud dependency, reducing latency, enhancing security, and lowering operational costs. With industrial-grade reliability, an M.2 Key M form factor, and operating temperatures from -40°C to 85°C, the Hailo-10H is ideal for smart cities, robotics, smart retail, transportation, and intelligent edge computing. The scalable UP x Hailo platform enables faster AI deployment, simplified development, and future-ready performance for next-generation embedded AI solutions.