Aetina’s New SuperEdge Product Series Designed for Machine Learning and Inference Applications

March 02, 2021 by Sara McCaslin

Aetina launches a new edge AI inference platform that is NVIDIA NGC-ready and scalable.

The Taiwanese company Aetina specializes in edge AI solutions for embedded applications and high-performance GPGPUs. The industries served by Aetina include aerospace, security, transportation, industrial automation, and defense.


The integration between hardware and software in the SuperEdge systems supports accelerated deployment of AI infrastructure and allows users to manage their edge devices more quickly without compromising security. Image used courtesy of Aetina


Aetina recently announced its latest product series, the Aetina SuperEdge system. This system is an NGC-Ready AI Inference Platform designed for edge AIoT (Artificial Intelligence of Things) environments and their fleet of devices. AIoT leverages the power of AI with an IoT infrastructure to support proactive, rather than reactive, responses to data and information.


Aetina SuperEdge

Aetina describes the SuperEdge system as a Multi-Access Edge Computing platform that seamlessly integrates high-performance Intel Xeon server microprocessors that are scalable with multi-GPU accelerators. 

The components and accelerators are combined in a hardware configuration that supports optimized AI workloads and inference at the edge. The Aetina SuperEdge platform aims to strengthen the connection between edge devices and edge AI and the connection between the edge and the cloud. 

In addition, the Aetina SuperEdge system has been tested and validated by NVIDIA to support NVIDIA NGC software related to inference, machine learning, and deep learning applications, resulting in faster machine learning training on edge solutions management platforms. 


Features and Capabilities

The pilot product for the Aetina SuperEdge system is the NVIDIA T4 GPU-based AIS-D422-A1. It is a barebone AI inference platform that supports up to 2x GPUs and has an Intel Xeon processor with eight cores and 16 threads, along with a 1TB NVME SSD. The AIS-D422-A1 includes an Intelligent Management Platform that enables users to remotely manage, build, and deploy their AI edge solutions, and security is provided through a TPM (Trusted Platform Module). 


The SuperEdge product line currently includes the AIS-D422-A1 and the AIS-D422-A113A2T1. Image used courtesy of Aetina


The AIS-D422-A113A2T1 is their full AI inference platform that, in addition to everything provided with the barebone AIS-D422-A1 platform, includes an NVIDIA T4 GPU accelerator. One of the benefits of the Aetina SuperEdge system is developing edge AI solutions more efficiently.

The integrated NVIDIA Tesla T4 GPU accelerator can achieve machine learning training up to 35X faster and deep learning training 10X faster when compared to CPU-only systems. 


AI on the Edge

Both the AIS-D422-A1 and AIS-D422-A113A2T1 are fully compatible with AI tools and software from the NGC catalog. This allows developers to configure and deploy GPU-optimized AI solutions far more quickly. 

The SuperEdge AI inference platform is unique to any industry that uses AI as part of the IoT (or IIoT) environment and wants to enhance the speed of both AI application development and machine learning. Note that the AIS-D422-A1 is available for purchase, and the AIS-D422-A113A2T1 will likely be released for purchase in the near future.