![\"New](\"https:\/\/scitechdaily.com\/images\/New-Light-Based-AI-Chip.jpg\")
<\/a>A brand new AI chip powered by gentle, not electrical energy, delivers large power financial savings whereas matching typical efficiency.<\/strong><\/p>\n A bunch of engineers has created a breakthrough laptop chip that depends on gentle fairly than electrical energy to hold out some of the energy-demanding capabilities in artificial intelligence<\/span>: identifying images and detecting patterns.<\/p>\n By shifting to light, the chip requires far less energy, achieving efficiencies that are 10 to 100 times greater than today\u2019s chips running the same types of calculations. This innovation could ease the massive strain AI places on power grids while also supporting the development of more advanced and capable AI models.<\/p>\n The key operation involved is known as \u201cconvolution,\u201d a process central to how AI interprets photos, videos, and even written language. At present, convolution is highly resource-intensive and time-consuming. The new design solves this by integrating lasers and tiny lenses directly onto circuit boards, allowing the chip to complete these computations with significantly less energy and greater speed.<\/p>\n In early trials, the chip achieved about 98% accuracy<\/span> when recognizing handwritten digits, matching the performance of conventional electronic chips.<\/p>\n \u201cPerforming a key machine learning<\/span> computation at near-zero energy is a leap forward for future AI systems,\u201d said study leader Volker J. Sorger, Ph.D., the Rhines Endowed Professor in Semiconductor Photonics at the University of Florida<\/span>. \u201cThis is critical to keep scaling up AI capabilities in years to come.\u201d<\/p>\n \u201cThis is the first time anyone has put this type of optical computation on a chip and applied it to an AI neural network,\u201d said Hangbo Yang, Ph.D., a research associate professor in Sorger\u2019s group at UF and co-author of the study.<\/p>\n Sorger\u2019s team collaborated with researchers at UF\u2019s Florida Semiconductor Institute, the University of California, Los Angeles and George Washington University on study. The team published their findings on September 8 in the journal Advanced Photonics<\/span>.<\/em><\/p>\n The prototype chip uses two sets of miniature Fresnel lenses using standard manufacturing processes. These two-dimensional versions of the same lenses found in lighthouses are just a fraction of the width of a human hair. Machine learning data, such as from an image or other pattern-recognition tasks, are converted into laser light on-chip and passed through the lenses. The results are then converted back into a digital signal to complete the AI task.<\/p>\n This lens-based convolution system is not only more computationally efficient, but it also reduces the computing time. Using light instead of electricity has other benefits, too. Sorger\u2019s group designed a chip that could use different colored lasers to process multiple data streams in parallel.<\/p>\n \u201cWe can have multiple wavelengths, or colors, of light passing through the lens at the same time,\u201d Yang said. \u201cThat\u2019s a key advantage of photonics.\u201d<\/p>\n Chip manufacturers, such as industry leader NVIDIA, already incorporate optical elements into other parts of their AI systems, which could make the addition of convolution lenses more seamless.<\/p>\n \u201cIn the near future, chip-based optics will become a key part of every AI chip we use daily,\u201d said Sorger, who is also deputy director for strategic initiatives at the Florida Semiconductor Institute. \u201cAnd optical AI computing is next.\u201d<\/p>\n Reference: \u201cNear-energy-free photonic Fourier transformation for convolution operation acceleration\u201d by Hangbo Yang, Nicola Peserico, Shurui Li, Xiaoxuan Ma, Russell L. T. Schwartz, Mostafa Hosseini, Aydin Babakhani, Chee Wei Wong, Puneet Gupta and Volker J. Sorger, 8 September 2025, Advanced Photonics<\/i>. Never miss a breakthrough: Join the SciTechDaily newsletter.<\/a><\/b><\/p>\n<\/p><\/div>\nA Leap Forward in AI Efficiency<\/h4>\n
Advantages of Optical Computing<\/h4>\n
DOI: 10.1117\/1.AP.7.5.056007<\/a><\/p>\n