DEEPX and Network Optix Join Forces: Faster and More Efficient Edge AI Processing

DEEPX has integrated its DX-M1 and DX-H1 NPU processors into the Nx EVOS platform using Network Optix’s AI Manager and the OAAX standard, enabling faster deployment, greater scalability, and high energy efficiency for edge AI systems. Thanks to their low power consumption and stable performance under heavy workloads, DEEPX accelerators offer significant advantages for video analytics and real-time data processing.

“The Nx EVOS platform provided an ideal environment for integrating our high-performance, energy-efficient NPUs. The combination of Nx’s open architecture and DEEPX technology sets a new standard for edge AI video analytics,” the DEEPX team stated.

The integration was carried out through a unified OAAX interface, allowing DEEPX engineers to quickly connect their SDK and runtime environment without additional hardware adjustments. Using Nx AI Manager, models can now be converted into the DEEPX dxnn format and distributed via the Nx Cloud platform across multiple servers and edge devices without manual configuration.

“The OAAX framework enabled us to efficiently port our SDK and model execution tools. The combination of the Nx EVOS cloud system and the OAAX standard made the model conversion and deployment process extremely fast and intuitive,” DEEPX added.

With an optimized SDK and energy-efficient NPU architecture, the new system delivers real-time inference with significantly lower power consumption and reduced CPU usage compared with traditional CPU- or GPU-based solutions. Close collaboration between DEEPX and Network Optix further improved stability and performance in demanding operational environments. “This project demonstrated strong synergy between our NPU optimization expertise and the compatibility of the Nx platform,” the DEEPX team concluded, noting that the integration serves as a model for other AI accelerator manufacturers seeking to join the Nx ecosystem.