TPU Inference Servers for Efficient Edge Data Centers - DOWNLOAD

Edge AI startup Femtosense to showcase AI silicon chip for audio applications at CES 2023

Categories Edge Computing News  |  Hardware
Edge AI startup Femtosense to showcase AI silicon chip for audio applications at CES 2023

Femtosense, a provider of edge AI infrastructure, has announced the successful testing of its first silicon product for audio applications. The company claims the chip will allow for more efficient real-time AI processing in consumer electronic products at a low cost. Femtosense designed the AI system-on-chip with Imec.IC-link, an ASIC solutions provider.

Femtosense utilizes near-memory compute technology to decrease the distance data must travel by performing computations close to on-chip memory banks. According to the company, this technique also helps to alleviate energy and memory bottlenecks caused by accessing off-chip memory. The scalable core design of Femtosense’s SPU architecture enables the digital design to be ported to other process nodes, allowing for flexibility in terms of performance and cost, the company says.

Femtosense will showcase wearable technology products, such as hearing aids and earbuds, that feature built-in noise suppression and speech enhancement technologies at the Consumer Electronics Show in 2023.

“We are already working with partners and customers to evaluate and develop products that enable unprecedented features for small-state battery-powered wearables down to home security, televisions and even audio in vehicles, all with AI that requires much less power for even greater efficiency,” said Sam Fok, Ph.D., CEO and co-founder, Femtosense.

Before this technology, Femtosense said providing AI programs outside data centers was cost-prohibitive and used significant energy. This company says its application of sparse mathematics in AI processing can conserve power and boost efficiency by reducing unneeded AI work.

“Femtosense and IC-link engaged in a very early stage of the ASIC development,” said Yiyi Wang, head of Imec.IC-link North America. “We are impressed with Femtosense’s design and engineering capabilities and how fast this team can deploy the design with all the features and turn the first silicon around within months.”

Last fall, Femtosense raised $8 million in Series A funding, led by Fine Structures Ventures, to support the development and delivery of the AI chips.

Article Topics

 |   |   |   |   |   | 

Comments

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Featured Edge Computing Company

Edge Ecosystem Videos

Latest News