The world's first USB edge AI accelerator — ASUS UGen300 with Hailo-10H chip delivering 40 TOPS (INT4) at just 2.5W, with 8GB LPDDR4 and a plug-and-play USB-C interface for LLM, VLM, and vision inference on any x86 or Arm host.
Hailo-10H
40 TOPS INT4 · 20 TOPS INT8
8GB LPDDR4 @ 4266 MT/s
USB 3.1 Gen2 Type-C (10 Gbps)
2.5W typical (5V via USB-C)
105 × 50 × 18 mm
145 g
x86, ARM
Windows, Linux, Android
100+ (LLM, VLM, Whisper, vision)
TensorFlow, TF Lite, Keras, PyTorch, ONNX
0°C to 40°C
72+ hours rated
Fanless — large internal heatsink
Available
Quote
The ASUS UGen300 is the world's first USB edge AI accelerator for both classic and generative AI. Powered by the Hailo-10H processor with 8GB of dedicated LPDDR4, it delivers up to 40 TOPS (INT4) of inference at a typical 2.5W — turning any laptop, edge computer, or industrial host into a secure, offline GenAI device through a standard USB-C port.
Plug-and-play across Windows, Linux, and Android on x86 or Arm, the UGen300 runs 100+ pre-trained models (LLMs, vision-language models, Whisper speech-to-text, and vision networks) via TensorFlow, PyTorch, ONNX, and Keras. Its slim fanless 105×50×18mm aluminum design with an internal heatsink makes it ideal for industrial, retail, robotics, and enthusiast deployments that need always-on local inference without a discrete GPU.
Need ASUS UGen300?
Enterprise hardware — configured, tested, deployed. Contact QS Compute for pricing, availability, and integration services.
Request Quote