Neuromorphic Photonic Computing: Lights On

By Anne-Françoise Pelé, EETimes Europe | November 26, 2025

As quantum and analog neuromorphic systems move closer to commercialization, new architectures continue to expand the definition of what computing can be.

“Usually, at CEA-Leti, research teams plan five years ahead, but in this case, it’s more like 10 years,” Éléonore Hardy, silicon photonics partnership manager at CEA-Leti, told EE Times Europe at the Silicon Photonics Workshop co-organized by CEA-Leti and Soitec alongside ECOC 2025 in Copenhagen, Denmark. “It’s a fascinating endeavor, and some of us researchers need to go beyond simply meeting the industry’s five-year demands and look to the future.”

Such long-horizon thinking aligns with a broader shift underway in computing. As EE Times highlighted in this year’s Silicon 100, unconventional computing—once a niche academic territory—is emerging as a category of real technological potential. It has come to describe systems that depart from electronic, transistor-based architectures and instead use physical, chemical, biological, or thermodynamic mechanisms for computation. Examples range from reversible computing, which recycles energy and reduces heat dissipation by making computational steps reversible, to molecular computing, which uses DNA, and thermodynamic approaches, which rely on entropy-driven behavior.

To read the full article, click here