Boost Your AI-Enabled Industrial and Consumer Devices with PSOCTM Edge

By Embedded Computing Design Staff

November 11, 2025

Edge AI is a key enabler for the IoT, offering significant market opportunity for intelligent devices, both in the industrial and consumer spaces. Before we dive deeper into this, let’s level set and define two terms that often mean different things to different people, namely Edge AI and intelligent devices.

Edge AI refers to the execution of AI and machine learning (ML) algorithms directly on endpoint devices rather than relying on cloud processing. This ability to handle inferencing locally lets the device analyze sensor data, make predictions, and trigger actions in real time without the latency, bandwidth, or privacy concerns of continuous Cloud connectivity.

Intelligent devices, in this context, are embedded systems equipped with sufficient compute resources, typically including an MCU or MPU with integrated AI acceleration, to process data locally and adapt their behavior based on learned models. In industrial settings, Edge AI typically involves machine learning, and enables predictive maintenance, anomaly detection, and autonomous control of machinery. In the consumer space, it can power voice assistants, gesture recognition, and personalized device responses.

Market Opportunities

The convergence of IoT and AI is creating vast opportunities across these industrial and consumer sectors. For example, industrial Edge AI enables smarter automation, condition-based monitoring, and real-time quality inspection, allowing factories to increase uptime and efficiency while reducing maintenance costs. In the consumer space, intelligent devices are reshaping user interaction through voice, gesture, and contextual awareness, enabling more intuitive, personalized experiences. From wearables and smart home products to industrial sensors and robotics, the ability to embed intelligence at the Edge transforms static devices into adaptive systems.

Edge AI maximizes these opportunities by allowing devices to act instantly on locally gathered data, independent of network latency or Cloud access. By processing information close to the source, it delivers faster, more reliable, and privacy-preserving responses, all critical elements for safety and real-time control in industrial systems and for seamless user experiences in consumer products. Furthermore, on-device intelligence minimizes bandwidth and cloud-processing costs while enhancing scalability across large deployments. With hardware platforms like Infineon’s PSOC™ Edge series of MCUs, developers can efficiently implement trained neural networks directly on embedded devices, bringing advanced inference and decision-making to even the smallest endpoints.

The PSOC™ Edge MCUs

Infineon’s PSOC™ Edge family of MCUs fits the bill for these applications as it combines high-performance processing, advanced security, and efficient AI acceleration to enable next-generation intelligent devices. Built on Arm Cortex®-M55 core and Ethos™-U55 AI/ML Neural Processing Unit (NPU), these devices deliver exceptional compute power for real-time inference at the Edge. They integrate DSP extensions, vector processing, and hardware support for neural networks, all within a low-power architecture that’s optimized for battery operation. The PSOC™ Edge platform also features a low-power domain driven by a Cortex®-M33 core paired with Infineon’s NNLite ideal for always-on operations and use cases such as wake work and keyword spotting. In addition to these high-performance yet low-power core offerings, PSOC™ Edge provides robust connectivity options, a secured enclave with advanced secure boot architecture with hardware root of trust, and seamless integration with RTOS and cloud ecosystems.

Four different versions of the PSOC™ Edge MCU are available, starting with the base E81. The E82 expands on the E81 by adding graphics capabilities with a 2.5D Graphics Processing Unit (GPU) and MIPI-DSI/DBI interfaces. The E83 integrates the advanced Ethos-U55 NPU as well as AI-based vision, voice and audio capabilities, and the E84, at the high end, bundles all of those features including graphics and advanced AI. And Infineon is just getting started here; they pledge to continue making significant investments both in hardware and software.

A Complete Development Environment

That said, the hardware is complimented by a complete development and enablement ecosystem. The platform is fully supported by ModusToolbox™, Infineon’s cross-platform design environment that simplifies configuration, firmware development, and integration of connectivity and ML components.

PSOC™ Edge is also supported by DEEPCRAFT™ Studio, providing access to optimized “ready models” that accelerate ML deployment across vision, voice, and sensor-based workloads. The integrated AI development environment accelerates the deployment of ML models by providing a unified toolchain that simplifies every stage of the Edge AI workflow, from data preparation and neural network training to model quantization, optimization, and deployment. Developers can import trained models from popular frameworks such as TensorFlow Lite or PyTorch, then automatically fine-tune them for the PSOC™ Edge hardware using DEEPCRAFT™ Model Converter tool. By leveraging the whole DEEPCRAFT™ AI Suite, developers can focus on application innovation rather than low-level integration.

Zephyr RTOS Support

Infineon’s PSOC™ Edge family now integrates seamlessly with the Zephyr® RTOS, giving developers access to a standardized, open-source software environment backed by a growing global ecosystem. Zephyr® is an open-source RTOS designed for resource-constrained devices. It offers a scalable kernel, modular architecture, and broad hardware support, suiting it for embedded and IoT applications requiring reliability, portability, and real-time responsiveness.

This integration simplifies firmware development and accelerates time-to-market by leveraging Zephyr®’s unified driver framework, connectivity stacks, and security modules. Developers can take advantage of Zephyr®’s modularity to efficiently manage multitasking, peripherals, and low-power modes across complex AI-enabled systems. Combined with PSOC™ Edge’s advanced compute and security architecture, Zephyr® enables a cohesive software foundation for scalable, production-ready embedded AI designs that are easily portable across future Infineon and third-party platforms.

To simplify design and integration, Infineon is offering two evaluation kits, which include BOMs and schematics, as well as lots of software. The kits also include the company’s PSOC 4000T CAPSENSE™ coprocessor for designs that require touch input. Wi-Fi, Bluetooth® and digital microphones are present on the kits as well.

Contact Infineon Technologies today to learn more.