The gist: Lightmatter, a photonic computing startup founded by MIT researchers in 2017, aims to overcome AI infrastructure bottlenecks by using light-based technology to enhance bandwidth, speed, and energy efficiency in data processing.
Lightmatter and the AI Infrastructure Challenge
Nicholas Harris is the founder and CEO of Lightmatter, a photonic computing startup established in 2017 by MIT researchers. The company addresses a critical bottleneck in AI infrastructure: conventional electronic chips and copper interconnects consume excessive power and restrict the speed of data transfer between processors.
Lightmatter’s innovative approach leverages photonics—computing and communication using light—to significantly improve AI system bandwidth, speed, and energy efficiency.
Relation to Reservoir Computing and Physical Systems
While Lightmatter is not strictly a reservoir-computing company, it aligns with the broader trend of employing physical systems instead of solely conventional digital logic to achieve more efficient computation. Both photonic computing and reservoir computing utilize physical dynamics to process complex information.
From the author’s perspective, particularly regarding memristor chips for biological applications, Lightmatter serves as a valuable example of how deep-tech hardware founders can attract investors by connecting advanced device physics to clear market demands.
Author's Personal Interest
The author is interested in memristor-based reservoir computing for processing neural signals, biosignals, or organoid electrophysiology in real time. This technology could enable low-power brain-computer interfaces (BCI) and closed-loop neuromodulation, illustrating how physical computing approaches can address complex biological data challenges.
Lightmatter exemplifies how linking advanced device physics with pressing market needs can attract investment and drive innovation in deep-tech hardware.