Embedded World 2025: GigaDevice Showcases Advances in Flash and MCU Technologies

GigaDevice, semiconductor company specializing in flash memory, 32-bit microcontrollers (MCUs), sensors, analog products and solutions, has announced it will attend Embedded World, in Nuremberg, Germany, and showcase, in Hall 5, Stand 5-129, its latest advances in flash and MCU technologies.

At the show, the company will demonstrate both the resilience and the adaptability of its solutions across 4 strategically important yet diverse applications: edge AI, industrial and green energy, automotive, and connectivity.

Serial NOR Flash for Automotive applications
The company‘s GD25/55 family of serial NOR flash has been certified to meet ISO 26262:2018 ASIL D automotive functional safety standards and strictly adheres to AEC-Q100 Grade 1. This 2Mb to 2Gb capacity flash memory offers 1-, 2-, 4- and 8-channel communication modes and a data throughput rate of 400MB/s – creating high-quality products to meet strict automotive application requirements.

The firm‘s 2nd-gen GD32A7 series automotive MCUs, built on a high-performance Arm Cortex-M7 core, offer enhanced computing power, security, and rich peripherals for applications such as body domain control, telematics, lighting, and battery management. Multiple demonstrations based on GD32A7 will be visible on stand, which includes secure communication, BDC, BMS and ADB.

Seamless smart home and IoT connectivity
GigaDevice’s GD32VW553 Matter solution utilizes a high-performance RISC-V core at 160MHz to enable secure data transmission in complex smart home and IoT applications. With support for Wi-Fi 6 (802.11 b/g/n/ax) and Bluetooth LE 5.2, the evaluation board integrates securely into Matter ecosystems, allowing devices to be controlled via the Tuya Smart App.

AI algorithms for a GD32H7 MCU-based neural net AFCI
The company‘s advanced Arc Fault Circuit Interrupter (AFCI) technology enables electrical safety in commercial and industrial settings by detecting dangerous arc faults in real time, preventing electrical fires. The demo, based on the GD32H7 with a 500K sampling rate and 12 ADC channels for real-time AI inference, utilizes a deep convolutional neural network processing 4,096 sampling points. It showcases precise detection of abnormal arcing while distinguishing harmless operations. Compatible with various hardware, the model enables one-click training for edge AI in adaptive function compute infrastructure. With cutting-edge detection and intelligent protection, the firm reaffirms its commitment to electrical safety innovation.

Keyword Recognition: Offline AI Solution for GD32W515 with Customizable Training and Minimal Memory Usage
The company‘s offline keyword-recognition solution, built on the GD32W515 reference design, utilizes an optimized AI algorithm that ensures minimal memory usage while delivering high-performance results. Companion software is also under development to allow the easy customization of keywords, and the ability to retrain it in less than 24 hours. The solution supports both training with and without target voice samples, offering flexibility for a wide range of applications. Additionally, the company is developing an online keyword detection feature for the GD32W515, which will further enhance its capabilities.

Real-time motion control in industrial and green energy systems
The GD32H75E enables real-time EtherCAT servo control for industrial automation, supporting the CiA 402 protocol and seamless TwinCAT integration. It delivers precise motion control through Profile Position mode (PP), 30 origin return Homing Modes (HM), and Cyclic Synchronous Position mode (CSP), with trapezoidal trajectory planning for smooth motion and zero-backlash modes for precise homing.

Bi-directional energy storage inverter
GigaDevice has created a bi-directional energy storage inverter based on its GD32G553 MCU. This system enables energy flow between batteries and the grid for enhanced power stability. It features a modular design with two MCUs controlling the three-layer (LLC, Buck/Boost, and inverter/PFC) architecture.