#India #Bangalore #Ambient Scientific #GPX10 Pro #AI processor

Ambient Scientific Unveils the GPX10 Pro: A New Era for Edge AI Processing

Ambient Scientific has recently launched its groundbreaking GPX10 Pro, a system-on-chip (SoC) designed specifically for battery-powered edge applications. With an impressive performance boost of up to 100 times compared to traditional 32-bit microcontrollers, this innovative chip is set to revolutionize the field of artificial intelligence at the edge.

Innovative Architecture

At the heart of the GPX10 Pro's capabilities lies its unique architecture, which utilizes AI-native silicon technology. The chip is equipped with 10 programmable AI cores that are optimized to handle various neural network models, such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), Long Short-Term Memory (LSTM), and Gated Recurrent Unit (GRU) models. Unlike conventional processors limited by their generalized instruction sets, the GPX10 Pro specifically tailors its architecture for AI tasks, allowing for efficient execution of complex processes like voice recognition, intelligent sensing, and computer vision.

Power Efficiency

One of the key highlights of the GPX10 Pro is its remarkable power efficiency. In certain scenarios, the chip operates on less than 100µW, which is particularly impressive considering it can maintain functionality while powered by a single coin-cell battery. This makes it an ideal choice for wearables and IoT devices that require high performance without consuming large amounts of power. The chip's architecture allows multiple operations to be carried out simultaneously, with the ability to perform over 2,560 multiply-accumulate operations per cycle, thereby achieving a peak AI throughput of 512 Giga Operations Per Second (GOPs).

Features and Functionalities

The GPX10 Pro isn't just a powerhouse in terms of computation; its design principles allow for a high level of integration. The chip is supported by 2MB of on-chip SRAM—ten times more than its predecessor, the GPX10—enabling it to handle larger and more complex AI models. Additionally, an integrated Arm® Cortex®-M4F CPU core facilitates classic control functions, while it also blends analog and digital functionalities seamlessly. Enhanced signal processing capabilities such as ultra-low power Analog-to-Digital Converters (ADCs) and interfaces for up to eight analog and twenty digital sensors make the GPX10 Pro a versatile option for various applications.

The Nebula Toolchain

To further enhance the chip’s appeal, Ambient Scientific provides the Nebula™ AI enablement toolchain, which simplifies the training, development, and deployment of AI models compatible with leading frameworks like TensorFlow, Keras, and ONNX. This software ecosystem allows developers to adapt their designs quickly as AI models evolve. Moreover, the onboard SenseMesh™ hardware sensor fusion layer offers ultra-low latency sensor fusion capabilities, enabling faster response times and lower idle power consumption by offloading sensor polling from the CPU.

Real-World Applications

At the upcoming Electronica event from September 17 to 19, 2025, in Bangalore, India, Ambient Scientific will showcase various real-time AI applications powered by the GPX10 Pro. Demonstrations will include fall detection, voice recognition, and facial recognition systems—all operating efficiently on a single coin-cell battery.

Availability

Currently, the GPX10 Pro is available for sampling, and mass production is anticipated to start in the first quarter of 2026. With its advancements in AI processing capabilities, Ambient Scientific aims to empower a range of devices, from tiny sensors to large-scale cloud infrastructures, paving the way for smarter, more efficient edge computing solutions.

In summary, the GPX10 Pro represents a significant leap forward in edge computing technology, delivering unmatched power efficiency and performance, poised to take the age of artificial intelligence to new heights.