Glass chips could be the answer to more computing power at the edge

Edge devices, including smart meters, smart home assistants, connected vehicles, and other IoT devices, rely primarily on the cloud for computing resource due to their small size and power limitations.

This means they need a constant data link to work effectively and that brings its own problems. Technology company Cognifiber has announced the development of a glass-based 'photonic chip' that has the potential to revolutionize edge computing.

"The downsizing potential using glass-based photonic chips in conjunction with our proprietary fibers promises to bring superb-performance servers to the edge, removing many existing bottlenecks while dramatically reducing power consumption," says Dr. Eyal Cohen, co-founder and CEO of Cognifiber. "Anything that generates vast amounts of data every second, such as connected vehicles, automated trains, or fleet management of large shipment drones can respond in real-time to events without reliance on data centers."

Future glass photonic chips, beyond downsizing, may provide a replacement for today's silicon ones, while having the potential for a 100-fold boost in computing capabilities and an 80 percent reduction in AI/ML training costs, as well as reducing manufacturing costs, power consumption, and removing bandwidth bottlenecks.

By replacing legacy silicon-based semiconductors the technology could redefine Moore's Law. The company is already in the advanced stages of developing in-fiber processing that minimizes the reliance on chips altogether by conducting complex computations within special optical fiber.

"The future of computing demands a whole new way of transferring and processing vast amounts of data," says Professor Ze'ev Zalevsky, co-founder and CTO of Cognifiber. "Combining photonic glass chips promotes our edge solution to bring rapid AI and Machine Learning locally to edge devices, which are limited in their capacity and power allowance."

You can find out more on the Cognifiber site.

Image credit: BeeBright/depositphotos.com