Hailo-10H M.2 — Edge AI Accelerator for Generative AI

Specifications

Overview

The Hailo-10H M.2 is the industry's first edge AI accelerator purpose-built for on-device generative AI, delivering 40 TOPS INT4 at just 2.5W typical power. Featuring Hailo's second-generation neural core architecture with up to 8GB on-module LPDDR4, it enables local inference of large language models (LLMs), vision-language models (VLMs), and real-time computer vision — without cloud dependency. The standard M.2 2242/2280 form factor with PCIe Gen3 x4 interface allows drop-in integration into existing edge PCs, industrial gateways, and embedded systems, backed by Hailo's Model Zoo and full TensorFlow/PyTorch/ONNX framework support.

Key Benefits

Generative AI at the Edge: First-to-market M.2 accelerator capable of running LLMs and VLMs locally — 8GB on-module RAM enables models previously requiring cloud or datacenter GPUs. 2.5W Power Envelope: Best-in-class power efficiency — 16 TOPS per watt — enables fanless, battery-powered, and thermally constrained deployments. Industrial Grade: -40°C to 85°C operating range with M.2 socket reliability for automotive, outdoor, and factory-floor AI at the edge. Drop-In Upgrade: Standard M.2 Key M slot with standard OS drivers — upgrade any x86 or ARM edge system with 40 TOPS in minutes.