ASUS Japan will release three models from the edge AI accelerator "ASUS UGen300" series and the "ASUS UGen200" on October 16. The UGen300 series features the Hailo-10H chipset and will be available in USB and M.2 connection models.
The UGen300 series combines the Hailo-10H, offering up to 40 TOPS of inference performance, with 8GB or 2GB of LPDDR4/4X memory. Its key feature is the ability to handle not only traditional vision AI but also LLM, VLM, and Whisper-based speech recognition without cloud dependency. ASUS offers the "S-UGEN300-8GLD4-TC," which simply plugs into a USB 3.2 Gen 2 Type-C port, and the "UGEN300-8GLD4-M2" and "UGEN300-2GLD4X-M2" for M.2 2280 slots, providing configurations that are easy to integrate depending on the application.
S-UGEN300-8GLD4-TC
The entry-level "UGen200-640KSR-M2" is equipped with the Hailo-8L chipset, offering up to 13 TOPS of performance, suitable for traditional AI processing such as object detection, image classification, and person tracking. The typical power consumption is 2.5W for the UGen300 series and 1.5W for the UGen200, making them easy to use in fanless enclosures, battery-powered devices, and environments requiring 24/7 operation. They widely support Windows, Linux, and Android operating systems, as well as x86 and ARM architectures, and come with over 150 pre-trained models and sample codes.
UGen200-640KSR-M2
The product lineup consists of four models: S-UGEN300-8GLD4-TC, UGEN300-8GLD4-M2, UGEN300-2GLD4X-M2, and UGen200-640KSR-M2. The USB connection is suitable for adding to existing PCs, while the M.2 connection is ideal for integration into custom-built PCs, industrial PCs, and single-board computers. ASUS positions these products as a lineup that meets the demands of development sites looking to achieve low-power AI inference in local environments, as well as applications in smart cameras, industrial sensing, and PoC (Proof of Concept) use cases.
