Aetina Corporation, which is an established manufacturer of Edge AI solutions, has revealed the launch of its next-gen MXM AI Accelerator Module series powered by NVIDIA RTX PRO Blackwell Embedded GPUs as a part of its CoreEdge MXM AI Module family. Built on the NVIDIA Blackwell GPU architecture, this new series offers up to 3X the AI performance uplift over its predecessor, setting a new standard for advanced AI computing at the edge.
The modules are designed to be compact and efficient as well as robust and are compliant to the MXM 3.1 industrial standard. They are purpose-built for high-precision, high-density workloads in the most critical applications like autonomous mobile robots – AMRs, unmanned vehicles – UAVs/UGVs, smart manufacturing equipment, as well as medical and surgical assistance systems. Generative AI and physical AI bring the computing power nearer to the edge, enabling an extensive upgrade of smart applications.
NVIDIA Blackwell Architecture – An Unmatched Performance Proven
As AI applications move from data centers to the edge, Aetina’s CoreEdge embedded MXM AI modules provide a small form factor and industrial-grade design with NVIDIA RTX PRO Blackwell GPUs, delivering cutting-edge AI and ray tracing, along with neural graphics technologies.
Aetina, driven by the NVIDIA RTX PRO 5000 Blackwell GPU, the flagship MX5000B-XA model, packs 10,496 CUDA cores, 80 RT cores, and 320 Tensor cores, delivering a performance of 40.62 TFLOPS FP32. Validation testing shows 2x faster LLM estimation on the Gemma-2-9b model and 80% faster AI inference performance in Geekbench AI benchmarks.
The NVIDIA Blackwell MXM series supports FP4 precision, DLSS 4 multi-frame generation technology, and GDDR7 memory, with bandwidth up to 896 GB/s and up to 3X higher AI computing throughput. These performance advancements enable the implementation of advanced AI models like large language models – LLMs and vision transformers in space- and power-constrained edge environments, resulting in lower latency and greater throughput computing when it comes to demanding applications like physical AI.
Industrial-Grade Durability Designed for Critical Environments
CoreEdge embedded MXM AI modules by Aetina are one-fifth the size of traditional graphics cards on the board but provide workstation-level computing performance. This Next-Gen MXM AI Accelerator Module Series design from Aetina driven by the NVIDIA Blackwell platform surpasses industrial standards, unlike commercial-grade products. It supports an extensive temperature range of -40°C to +85°C, with dust and moisture-resistant conformal coating protection. It also features configurable TGP power settings and high-vibration-resistant mounting holes on MXM Type B modules for secure installation in high-vibration environments. Aetina also offers 5-year product lifecycle support for supply stability in the long run.
This high reliability comes from Aetina’s rigorous engineering process, which includes wide-temperature components that have high-reliability packaging, thorough thermal cycling testing, cold boot inspection, and 100% chamber testing before shipment. This allows system integrators and developers to deploy with full confidence in mission-critical applications from autonomous robots and unmanned vehicles to smart manufacturing and medical equipment. Aetina also offers evaluation kits with PCIe Gen4 x16 interfaces and HDMI outputs to speed up development processes and ease integration.
Generative and Physical AI at the Edge – Putting Forward
According to the Director of the AI Accelerator Card & GPU Products Division at Aetina, Jackal Chen, “Aetina’s CoreEdge embedded MXM AI accelerator module series demonstrates our commitment to delivering exceptional performance and reliability. By incorporating NVIDIA Blackwell technology alongside Aetina’s extensive engineering expertise, we achieve high-performance AI computing for even the most demanding environments. Additionally, we are prepared for the age of AI agents, multimodal applications, and real-time AI-accelerated computing. We offer comprehensive value-added engineering services to streamline system integration processes, helping customers significantly reduce their time-to-market and fully harness the potential of edge generative AI and physical AI.”