Latest AI Inference Platform for AIoT Applications

Aetina, Taiwan, 15.12.2022 - The AIP-SQ67 series is a powerful AI inference platform which brings unrivalled performance and expandability for use in AI and IoT. Unlike AI model training, AI inference doesn’t re-evaluate or adjust the neural network based on the results. It simply applies knowledge from a trained model and outputs a prediction. An ideal AI inference system requires high computing performance to run the trained AI model, and low latency during the inference process to make decisions in real time. High I/O bandwidth and enough memory are both vital for the inference process.

The AIP-SQ67 series is designed and ready to help deliver the necessary bandwidth to execute advanced AI algorithms. It’s compatibility with different types of sensors and expansion modules makes it a great fit for most complex AI inference applications, including factory automation, smart cities, retail, healthcare, logistics, and transportation applications.

Premium Performance, Highly Compatible

Each AIP-SQ67 platform features a 12th Gen Intel® Core™ i5/i7 processor which provides superior processing performance. To deliver latency-free and high-bandwidth storage experience, the AIP-SQ67 series is compatible with one MXM module and up to two ASIC-based M.2 M-key modules supporting PCIe Gen4x4 and NVME. Even more, its built-in out-of-band (OOB) remote management module allows system operators to manage and maintain devices even when a device’s hardware or software is malfunctioning.

Rich I/O Interface for Diverse Peripherals and Flexible Expansion

The AIP-SQ67 series offers a variety of I/O interfaces, including 5 x DisplayPort (DP++) ports, 5 x 2.5G RJ45 LAN ports, 6 x USB 3.2 Gen2 x 1 Type-A ports, and 1 x USB 3.2 Gen2 x 2 Type-C port, allowing developers to integrate different, multiple sensors, such as 4K cameras, for image or video analytics tasks. It also features two swappable hard drive bays for 2 x 2.5" SATA SSD/HDD, enables users to handle their maintenance tasks with less effort.