Aalto University researchers performed AI tensor operations with a single pass of light, encoding data into light waves for passive, simultaneous calculations integrated into photonic chips for faster, energy-efficient AI systems. The study was published in Nature Photonics on November 14th, 2025.
Tensor operations, crucial for AI in image processing and language understanding, are advanced mathematical computations. Conventional digital hardware, including GPUs, faces speed, energy use, and scalability challenges with increasing data volumes.
An international team, led by Dr. Yufeng Zhang from Aalto University’s Photonics Group, developed an approach enabling complex tensor calculations in a single movement of light through an optical system. This process, termed single-shot tensor computing, operates at the speed of light.
“Our method performs the same kinds of operations that today’s GPUs handle, like convolutions and attention layers, but does them all at the speed of light,” stated Dr. Zhang. “Instead of relying on electronic circuits, we use the physical properties of light to perform many computations simultaneously.”
The team embedded digital information into the amplitude and phase of light waves, converting numerical data into physical variations within the optical field. These light waves interact, automatically performing mathematical procedures like matrix and tensor multiplication, fundamental to deep learning. Utilizing multiple wavelengths of light expanded the technique to support higher-order tensor operations.
“Imagine you’re a customs officer who must inspect every parcel through multiple machines with different functions and then sort them into the right bins,” Zhang explained. “Normally, you’d process each parcel one by one. Our optical computing method merges all parcels and all machines together — we create multiple ‘optical hooks’ that connect each input to its correct output. With just one operation, one pass of light, all inspections and sorting happen instantly and in parallel.”
Operations occur as light travels, eliminating the need for active control or electronic switching during computation.
“This approach can be implemented on almost any optical platform,” said Professor Zhipei Sun, leader of Aalto University’s Photonics Group. “In the future, we plan to integrate this computational framework directly onto photonic chips, enabling light-based processors to perform complex AI tasks with extremely low power consumption.”
Zhang indicated the goal is to adapt the technique to existing hardware and platforms used by major technology companies, estimating integration within 3 to 5 years.
“This will create a new generation of optical computing systems, significantly accelerating complex AI tasks across a myriad of fields,” he concluded.
