Lightmatter Unveils Breakthrough Photonic Chip to Boost AI Speed and Efficiency
Lightmatter, a technology startup located in Silicon Valley, has developed an advanced computer chip that promises to speed up artificial intelligence (AI) computations without increasing power usage as reported by Reuters.
The firm, which just hit a valuation of $4.4 billion after raising $850 million in venture capital, is one of a rising tide of companies looking into photonics by using light beams rather than electrical signals to move data between machines at faster speeds. The technology is vital for AI systems, which normally need huge networks of computers to handle complex workloads.
Pioneering Light-Based Computing
Beyond data transmission, Lightmatter is taking photonics a step further by using light beams to perform the calculations themselves. In a paper published in the journal Nature, the company outlined how its chip architecture diverges from traditional computing, which relies on ever-smaller transistors to increase performance.
According to Reuters, as shrinking transistors has become increasingly difficult, Lightmatter's solution bypasses this limitation by directing finely tuned beams of light to interact, then captures the results using integrated chips produced by manufacturing partner GlobalFoundries.
Also read: Google Unveils Ironwood as Next-Gen TPU for AI Inference
Overcoming Precision Challenges in Photonic Computing
Nick Harris, Lightmatter’s CEO, shared with Reuters that the chip can already handle certain AI tasks with accuracy comparable to traditional semiconductor chips. However, he acknowledged that it may take around a decade before the technology becomes widely adopted.
Harris said, “What we're doing is looking at the future of where processors can go. We fundamentally care about computers, and this is one of the alternative paths. Trillions of dollars of economic 's behind the idea that computers will keep getting better.”
Historically, photonic chips have struggled with precision, often losing accuracy when handling very small numerical values, which would sometimes be misread as zero. Lightmatter has tackled this issue by segmenting large and small numbers into groups before processing them through its photonic circuits. This approach helps preserve accuracy, even with delicate calculations.
