An AI‑optimized architecture with performance of
up to 40 eTOPS

AI workloads at the edge are rapidly increasing due to the rise of larger, multimodal models and on‑device AI processing. Meeting these growing demands requires more than conventional compute improvements. Advanced edge systems need dedicated acceleration at their core, delivering real‑time performance while maintaining energy efficiency, scalability, and strong data security.

For these requirements, the NXP Ara240 Discrete Neural Processing Unit (DNPU) delivers an AI‑optimized architecture with performance of up to 40 eTOPS. Featuring large on‑chip memory and high off‑chip bandwidth, the Ara240 is purpose‑built for advanced edge AI workloads such as large language models, vision‑language models, multimodal inference, and next‑generation AI applications.

Building edge AI solutions with a comprehensive design-in partner

Batenburg Applied Technologies is expanding its portfolio with two solutions based on this NXP technology, developed and manufactured by its long‑standing European partner F&S Elektronik Systeme. Designed and produced to the highest quality standards, F&S integrates the Ara240 DNPU into both a stand‑alone M.2 module and an i.MX95‑based SMARC module, enabling flexible deployment.

Both solutions are designed for demanding edge AI tasks, including industrial automation, autonomous robotics, smart infrastructure, advanced human‑machine interfaces, and edge servers. Powered by the Ara240 DNPU, device manufacturers gain the performance headroom required to run modern AI workloads locally at the edge, while ensuring scalable and future‑ready system designs.

Expertise focused on future-proof automation hardware

Through a hands‑on approach spanning the selection, design‑in, and optimization of embedded hardware and peripherals, Batenburg supports customers throughout the development of application‑oriented automation hardware. Backed by technical guidance, in‑house prototyping, system assembly expertise, and a strong focus on co‑creation, device manufacturers can accelerate time‑to‑market without compromising scalability, long‑term availability, or solution longevity.