Overview

• Lightmatter and Lightelligence have introduced operational photonic chips that perform matrix multiplication using light, a core function in AI processing.
• These chips integrate photonic and electronic components, achieving faster computations and significantly reduced energy consumption compared to traditional processors.
• The technology addresses the stagnation of Moore’s Law by circumventing the limitations of shrinking transistors in electronic chips.
• Manufactured using existing semiconductor production processes, the chips are scalable and could see integration in data centers within five years.
• While promising, experts predict widespread adoption of photonic computing may take up to a decade due to the need for further optimization and industry adaptation.