Designed for autonomous robotics, manufacturing tools, and medical units, the brand new modules present industrial-grade reliability and next-generation AI computing effectivity on the edge.
Aetina Company, a number one supplier of Edge AI options, has introduced the launch of its next-generation MXM AI Accelerator Module collection accelerated by NVIDIA RTX PRO Blackwell Embedded GPUs, which is a part of its CoreEdge MXM AI Module household. This new collection is accelerated by the NVIDIA Blackwell GPU platform to ship as much as 3X the AI efficiency uplift of its predecessor, establishing a brand new benchmark for superior AI computing on the edge
These modules are designed to be compact, environment friendly, and rugged, adhering to the MXM 3.1 industrial customary. They’re purpose-designed for high-precision, high-density workloads in crucial purposes akin to autonomous cellular robots (AMRs), unmanned automobiles (UAVs/UGVs), good manufacturing tools, and medical and surgical help methods. This collection brings the computing energy of generative AI and bodily AI nearer to the sting, facilitating a complete improve of good purposes.
NVIDIA Blackwell Structure: Confirmed Distinctive Efficiency
As AI purposes shift from knowledge facilities to edge environments, Aetina’s CoreEdge embedded MXM AI modules provide compact and industrially dependable designs that includes NVIDIA RTX PRO Blackwell GPUs, integrating breakthrough AI, ray tracing, and neural graphics applied sciences.
Aetina’s flagship MX5000B-XA mannequin, accelerated by NVIDIA RTX PRO 5000 Blackwell GPU, boasts 10,496 CUDA cores, 80 RT cores, and 320 Tensor cores, delivers a powerful 40.62 TFLOPS FP32 efficiency. Validation testing demonstrates 2x sooner LLM inference on the Gemma-2-9b mannequin and 80% sooner AI inference efficiency in Geekbench AI benchmarks.
By supporting FP4 precision, DLSS 4 multi-frame era expertise, and GDDR7 reminiscence with bandwidth as much as 896 GB/s, the NVIDIA Blackwell MXM collection boosts AI computing throughput by as much as 3X. These efficiency breakthroughs allow deployment of superior AI fashions akin to giant language fashions (LLMs) and imaginative and prescient transformers in space- and power-constrained edge environments, offering decrease latency and better throughput computing for demanding purposes akin to bodily AI.
Industrial-Grade Reliability: Engineered for Vital Environments
Aetina’s CoreEdge embedded MXM AI modules occupy solely one-fifth of the board area required by conventional graphics playing cards whereas delivering efficiency similar to workstation-class computing. In contrast to commercial-grade merchandise, Aetina’s collection design powered by the NVIDIA Blackwell platform surpasses industrial requirements. It helps a large temperature vary of -40°C to +85°C and options dust- and moisture-resistant conformal coating safety. Moreover, it presents adjustable TGP energy settings and high-vibration resistant mounting holes on MXM Kind B modules, making certain safe set up in high-vibration environments. Aetina additionally offers a 5-year product lifecycle help, guaranteeing long-term provide stability.
This complete reliability stems from Aetina’s rigorous engineering methodology, together with using wide-temperature elements with high-reliability packaging, in depth thermal biking testing, chilly boot testing, and 100% chamber testing earlier than cargo. This ensures system integrators and builders can deploy with absolute confidence in crucial purposes from autonomous robots, unmanned automobiles, good manufacturing and medical tools. Moreover, Aetina offers analysis kits outfitted with PCIe Gen4 x16 interfaces and HDMI outputs to speed up improvement processes and simplify integration.
Unleashing the Generative and Bodily AI on the Edge
“Aetina’s CoreEdge embedded MXM AI accelerator module collection demonstrates our dedication to delivering distinctive efficiency and reliability,” stated Jackal Chen, Director of AI Accelerator Card & GPU Merchandise Division at Aetina. “By incorporating NVIDIA Blackwell expertise alongside Aetina’s in depth engineering experience, we obtain high-performance AI computing for even essentially the most demanding environments, Moreover, we’re ready for the age of AI brokers, multimodal purposes, and real-time AI-accelerated computing. We provide complete value-added engineering providers to streamline system integration processes, serving to clients considerably scale back their time-to-market and absolutely harness the potential of edge generative AI and bodily AI.”
The MX5000B-XA and MX500B-QA can be found now. The MX4000B and MX2000B modules are anticipated to enter mass manufacturing in Q1 2026. For extra info, please go to www.aetina.com
About Aetina
Aetina, a supplier of state-of-the-art AI options, presents a variety of AI computing methods, platforms, {hardware}, and software program instruments ultimate for the creation of several types of vertical AI. Aetina’s built-in options empower its Arm, x86 computer systems, GPUs, ASIC {hardware}, cloud administration software program, and improvement instruments with synthetic intelligence, delivering complete, tailored {hardware} and software program suites to allow clean, fast system intellectualization progress on the edge. With its ecosystem community, Aetina makes the creation and deployment of any particular AI-powered purposes extremely achievable by leveraging its international companions’ AI applied sciences. Learn extra at: www.aetina.com
