A team of engineers from the University of Pennsylvania has developed a groundbreaking technology that revolutionizes the functionality of computer chips. The team has created a novel type of AI chip capable of utilizing light waves instead of electricity for computation, offering the potential for enhanced speed and cost savings.
This innovative AI chip is built upon the principles of silicon photonics (SiPh), which merges silicon, the predominant material used in computer chip fabrication, with the remarkable properties of light wave manipulation. This convergence is the result of pioneering research spearheaded by Penn Professor Nader Engheta. Professor Engheta’s work focuses on refining materials at the nanoscale to perform intricate mathematical operations using light waves.
READ MORE — Brillanz Group to Present AI-Powered Energy Intelligence Platform “Thunder” at MWC Barcelona Summit
Unlike traditional silicon chipsets, this new chip represents a potential paradigm shift in computing. While conventional chips have relied on electricity since their inception in the 1960s, the new chip operates on an entirely novel principle, performing calculations at the speed of light.
Firooz Aflatouni, Associate Professor in Electrical and Systems Engineering at Penn, played a significant role in this project. The team prioritized the execution of vector-matrix multiplication—a foundational process supporting neural networks crucial for modern AI applications.
The innovative design of this AI chip allows for the manipulation of silicon wafer height in specific areas, enabling “computational magic” by orchestrating light scattering patterns to expedite complex mathematical operations.
In addition to offering higher processing speeds and cost savings, this chip enhances privacy by performing calculations without storing data in the computer’s memory, making systems employing this technology highly secure against hacking attempts.
However, the commercial or industrial adoption of this cutting-edge technology is contingent upon achieving mass production, a process that may take considerable time.