In a major technological leap announced on June 20, 2025, two European research teams have successfully demonstrated how intense laser pulses through ultra-thin glass fibers can perform AI-like computations thousands of times faster than conventional electronic methods.
The breakthrough, led by Professors Goëry Genty from Tampere University and John Dudley and Daniel Brunner from the Université Marie et Louis Pasteur, represents a fundamental shift in computing architecture. By harnessing light instead of electricity for computations, the technology promises not only unprecedented processing speeds but also significantly improved energy efficiency.
"This work demonstrates how fundamental research in nonlinear fiber optics can drive new approaches to computation. By merging physics and machine learning, we are opening new paths toward ultrafast and energy-efficient AI hardware," explained the research team leaders. Their system achieved over 91% accuracy in classifying handwritten digits in under one picosecond—performance comparable to state-of-the-art digital methods but at vastly higher speeds.
What makes this approach particularly promising is that it doesn't simply rely on brute force. The researchers discovered that optimal results came from a careful balance between fiber length, dispersion (the propagation speed difference between wavelengths), and power levels rather than maximum nonlinear interaction.
Meanwhile, in a parallel breakthrough, AI researchers at Switzerland's Paul Scherrer Institute have developed a machine learning system that can formulate climate-friendly cement recipes in seconds. Led by mathematician Romana Boiger, the team created what they describe as "a digital cookbook for climate-friendly cement" that simulates and optimizes cement formulations to significantly reduce CO₂ emissions while maintaining structural performance.
This development addresses a critical environmental challenge, as cement production accounts for approximately 6% of global greenhouse gas emissions. The AI system can rapidly evaluate thousands of potential ingredient combinations, identifying formulations that reduce the carbon footprint while preserving cement's essential properties.
Together, these breakthroughs highlight how AI is both transforming computing infrastructure and helping solve pressing environmental challenges, pointing toward a future of faster, more efficient, and potentially more sustainable technological development.