Next-Gen Custom Chip Platform Boosts Battery Power, AI, and Edge Performance

The SFA 100, developed by Sondrel, features an Arm® subsystem with a variety of Arm cores and the option of a security subsystem. Its low power consumption makes it an excellent choice for battery-powered edge AI applications like smart homes, smart factories, and sensor fusion devices. This device offers the capability to integrate a machine-learning engine that supports various inference-based functionalities such as voice activation, image classification, gesture recognition, and filtering. Additionally, a radio subsystem like a Bluetooth IP block or WiFi can be incorporated if needed. Sondrel provides the flexibility to include customer IP and third-party IP by wrapping IP blocks with all the necessary interfaces for easy integration into the system. For more robust solutions, starting points like the SFA 200 or SFA 300 can be utilized.

John Chubb, the CEO of Sondrel, highlighted the growing trend towards AI integration in products at this year's CES. He emphasized the challenges of incorporating AI into devices, noting that it requires more powerful processor cores to handle the substantial amounts of data involved. Designing for AI from the ground up is essential to ensure that the processors, memory, DSPs, NoC, and other components align with the target Power, Performance, Area, and Timing specifications. Sondrel's approach, known as "Architecting the Future," streamlines the design process by leveraging existing blocks and layouts. By customizing pre-defined architectures like the SFA 100, Sondrel can efficiently create bespoke solutions for each client, reducing time to market and mitigating risks associated with new designs.

The Edge AI market is poised for significant growth, with a projected CAGR of 33.5% to reach $76 billion by 2031, as reported by Transparency Market Research. The shift towards processing data locally is a key driver in this expansion, aiming to eliminate latency by avoiding the need to transmit data to centralized cloud servers for processing. Edge AI applications span various sectors, including industry 4.0, where real-time decision-making is crucial, as well as consumer devices like drones, smartphones, and automobiles.

For more information on Sondrel's innovative solutions and how the SFA 100 platform can empower edge AI applications, visit www.sondrel.com. Stay ahead of the curve in the rapidly evolving landscape of AI technology with Sondrel's cutting-edge designs and expertise.