33 Views

Brainchip Launches Ultra-Low Power NPU in Market

October 01, 2024

Get a Price Quote

The Akida Pico NPU is making waves in the world of artificial intelligence with its claim to be the lowest power acceleration co-processor available. This innovative technology enables the development of compact, ultra-low power, portable devices that are intelligent enough to be integrated into various applications, including wearable technology, healthcare, consumer electronics, IoT, defense, and wake-up systems.

One of the key features of the Akida Pico is its ability to accelerate limited use case-specific neural network models, resulting in an ultra-energy efficient, purely digital architecture. This Neural Processing Unit (NPU) allows for secure personalization in a wide range of applications such as voice wake detection, keyword spotting, speech noise reduction, audio enhancement, presence detection, personal voice assistants, automatic doorbells, wearable AI, and appliance voice interfaces.

Powered by the Akida2 event-based computing platform configuration engine, the Akida Pico operates with remarkable energy efficiency, consuming less than a single milliwatt of power. This makes it an ideal solution for devices that rely on battery power, providing a power-efficient footprint for waking up microcontrollers or larger system processors. The NPU includes a neural network that filters out false alarms, ensuring power is preserved until an event is detected, making it perfect for sensor hubs or systems that require continuous monitoring with occasional processing needs.

Developers can leverage BrainChip’s exclusive MetaTF™ software flow to compile and optimize their Temporal-Enabled Neural Networks (TENNs) on the Akida Pico. With support for models created using TensorFlow/Keras and Pytorch, users can quickly develop and deploy AI applications for the Edge without the need to learn a new machine language framework.

Sean Hehir, CEO at BrainChip, highlighted the benefits of the Akida Pico, stating, “Like all of our Edge AI enablement platforms, Akida Pico was developed to further push the limits of AI on-chip compute with low latency and low power required of neural applications. Whether you have limited AI expertise or are an expert at developing AI models and applications, Akida Pico and the Akida Development Platform provides users with the ability to create, train, and test the most power and memory efficient temporal-event based neural networks quicker and more reliably.”

BrainChip’s Akida platform is designed for event-based computing, making it ideal for early detection and low-latency use cases in applications such as robotics, drones, automotive systems, and traditional sense-detect-classify-track systems. With a range of software, hardware, and IP products available for integration into existing and future designs, BrainChip is paving the way for customers to deploy multi-modal AI models at the edge.

For more information on BrainChip and the Akida Pico NPU, visit www.brainchip.com.

Recent Stories