Feature: System design
Te Innodisk Apex Series is a good
The challenge of pushing AI to the edge is not solely a hardware problem, it’s a question of development
example, the APEX-X100 uses NVIDIA RTX 6000 Ada accelerators, making it suitable for high-precision medical imaging and high performance computing (HPC) applications. Its durable design, which includes wide temperature support and resistance to vibration, ensures these solutions are reliable enough for use in harsh industrial environments and smart city infrastructure.
Mission-proven hardware Te imperative for power and reliability extends into the most challenging of domains – that of mission-critical environments. For defence, aerospace and space, standard embedded computing is not enough. Aitech Systems, a long- established developer of sturdy embedded systems, builds rugged GPGPU-based AI supercomputers on Nvidia’s Orin architecture. Te space-qualified S-A2300, for instance, is designed for low earth orbit missions, delivering the computational power required for autonomous satellite operations. In the air, their A178-AV system integrates with avionic platforms to enhance navigation and pilot situational awareness. Tese systems are engineered
to withstand extreme temperature fluctuations, shock and vibration, ensuring operational integrity in the most challenging environments.
Streamlining development for autonomy Te challenge of pushing AI to the edge is not solely a hardware problem, it’s a question of development. Te process of transitioning AI models to embedded systems has been complex and prone to vendor lock-in. A new focus on open, vendor-agnostic platforms is beginning to streamline this process, making powerful
www.electronicsworld.co.uk November 2025 21
AI development accessible to a wider range of engineers and businesses. Te Alp Lab Edge-1 AI Module (E1M) embodies this principle. Tis unified, plug-and-play module features a standardised pinout and a single soſtware stack (the Alp SDK) that eliminates traditional vendor dependency. Developers can move between different processor architectures – from Alif Semi to Renesas and Qualcomm – using the same hardware design and user code. Tis approach greatly reduces development time and costs, liberating engineers from vendor lock-in and allowing them to focus on innovation.
example of this, as it is based on a modular architecture that allows extensive customisation. Te series offers a comprehensive ecosystem of industrial- grade SSDs and DRAM modules, ensuring a cohesive and integrated solution for complex industrial deployments and smart city infrastructure. Te evolution continues with
microcontrollers that deliver generative AI in power-constrained environments. Alif Semiconductor’s Ensemble GenAI products are purpose built with an AI-ready architecture, featuring the Arm Ethos-U85 NPU, which supports transformer-based ML networks. As an example of its efficiency, an E4 device draws only 36mW when generating text, showcasing the power efficiency needed for applications in robotics and smart city equipment. It is clear that the future of AI is not a
single, centralised entity, but a distributed, intelligent network.
Page 1 |
Page 2 |
Page 3 |
Page 4 |
Page 5 |
Page 6 |
Page 7 |
Page 8 |
Page 9 |
Page 10 |
Page 11 |
Page 12 |
Page 13 |
Page 14 |
Page 15 |
Page 16 |
Page 17 |
Page 18 |
Page 19 |
Page 20 |
Page 21 |
Page 22 |
Page 23 |
Page 24 |
Page 25 |
Page 26 |
Page 27 |
Page 28 |
Page 29 |
Page 30 |
Page 31 |
Page 32 |
Page 33 |
Page 34 |
Page 35 |
Page 36 |
Page 37 |
Page 38 |
Page 39 |
Page 40 |
Page 41 |
Page 42 |
Page 43 |
Page 44
