Aetina, a leading provider of edge AI solutions, launched the AIB-MR1A-A1 AI Inference Platform from its DeviceEdge Series, marking the first edition of an ARM-based architecture with an AI module. This platform, powered by the Rockchip RK3588 ARM processor combined with the M.2 2280 AI accelerator, can manage up to 32 live or pre-recorded video streams. It is especially effective for security monitoring in large venues, offering cost-effectiveness, flexibility, and superior performance for advanced AI applications across various industries.
The rise of edge computing and generative AI has heightened the demand for ARM processors, known for their lower power consumption and total cost of ownership (TCO). These processors often struggle to keep pace with the rapid iteration of AI Processing Units (NPU), which can be time-consuming and costly. To meet these challenges, Aetina’s AIB-MR1A-A1 incorporates the Rockchip 8-core 64-bit ARM processor RK3588 with a PCIe Gen3 x4 M.2 2280 M-Key slot that supports the AI accelerators. This configuration significantly enhances AI computing acceleration and can manages 32 live video streams at 1080P, 30 FPS, delivering exceptional energy efficiency in video and image processing, including real-time video analytics and 8K video encoding/decoding.
Moreover, Aetina strategically collaborates with AI processor partners DEEPX and Hailo, providing a variety of M.2 AI accelerator options. This flexibility allows developers to easily swap AI module without needing to modify the hardware platform, drastically reducing time-to-market and increasing design adaptability.
Also Read: Houlihan Lokey Expands Global Technology Group With Senior Software Hire
Jackal Chen, Senior Product Manager at Aetina, stated, “The demand for flexible and energy-efficient ARM edge devices with AI accelerators is rapidly escalating. The AIB-MR1A-A1 addresses this demand by providing a market-ready industrial-grade AI platform and customized services that depart from traditional x86 architectures, enabling diverse edge AI computing applications.”
The AIB-MR1A-A1 can boost AI inference performance up to 26 TOPS through the integration of an AI accelerator. It also features multiple networking options and flexible I/O interfaces, supporting various peripheral and add-on modules like SATA storage. In addition, Aetina offers comprehensive customization services from chip-down design, BSP customization and thermal design to wide temperature testing and software integration.
SOURCE: Aetina