Tensor operations are a form of advanced mathematics that support many modern technologies, especially artificial intelligence. These operations go far beyond the simple calculations most people encounter. A helpful way to picture them is to imagine manipulating a Rubik’s cube in several dimensions at once by rotating, slicing, or rearranging its layers. Humans and traditional computers must break these tasks into sequences, but light can perform all of them at the same time. Today, tensor operations are essential for AI systems involved in image processing, language understanding, and countless other tasks. As the amount of data continues to grow, conventional digital hardware such as GPUs faces increasing strain in speed, energy use, and scalability.

The relentless demand for more powerful and efficient AI has pushed the boundaries of conventional computing. Graphics Processing Units (GPUs), which have become the workhorses of AI development, are reaching their physical and energetic limits. The intricate computations required for deep learning, such as convolutions and attention mechanisms, demand massive parallel processing. However, as datasets balloon and model complexity increases, GPUs struggle to keep pace, leading to bottlenecks in training times, exorbitant energy consumption, and challenges in scaling to meet future needs. This is where the innovative approach developed by an international team, spearheaded by Dr. Yufeng Zhang of Aalto University’s Photonics Group, offers a paradigm shift.

Researchers Demonstrate Single-Shot Tensor Computing With Light

To address these formidable challenges, an international team led by Dr. Yufeng Zhang from the Photonics Group at Aalto University’s Department of Electronics and Nanoengineering has developed a fundamentally new approach. Their method allows complex tensor calculations to be completed within a single movement of light through an optical system. The process, described as single-shot tensor computing, functions 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," says Dr. Zhang. "Instead of relying on electronic circuits, we use the physical properties of light to perform many computations simultaneously."

Encoding Information Into Light for High-Speed Computation

The ingenuity of this breakthrough lies in its ability to harness light’s natural wave-like properties for computation. The team accomplished this by embedding digital information into the amplitude and phase of light waves, transforming numerical data into physical variations within the optical field. As these light waves interact within a specially designed optical system, they automatically carry out mathematical procedures such as matrix and tensor multiplication, which form the fundamental building blocks of deep learning algorithms. This parallel processing capability is inherent to the nature of light. By strategically employing multiple wavelengths of light, the researchers have ingeniously expanded their technique to support even more complex, higher-order tensor operations, mirroring the increasing sophistication of AI models.

The analogy provided by Dr. Zhang vividly illustrates the transformative power of this optical computing approach. "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 says. "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." This "single-shot" operation means that a complex series of calculations that would take traditional computers many sequential steps is performed in a single, instantaneous event as light traverses the optical circuit.

Passive Optical Processing With Wide Compatibility

One of the most striking benefits of this method is its inherent simplicity and efficiency, requiring minimal external intervention. The necessary operations occur on their own as the light travels through the optical system, a phenomenon known as passive optical processing. This means the system does not need active control or complex electronic switching during the computation itself, dramatically reducing energy consumption and potential points of failure. "This approach can be implemented on almost any optical platform," says 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." This integration onto photonic chips signifies a crucial step towards practical, widespread adoption.

Path Toward Future Light-Based AI Hardware

The ultimate objective for Dr. Zhang and his team is to bridge the gap between their innovative research and the existing technological infrastructure of major technology companies. They aim to adapt this light-based computational framework to be compatible with the hardware and platforms currently employed in the AI industry. Zhang estimates that the method could be incorporated into such systems within a realistic timeframe of 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 concludes, envisioning a future where AI applications in areas like drug discovery, climate modeling, autonomous systems, and advanced scientific research are no longer constrained by computational limitations.

The study, which details this revolutionary advancement, was published in the prestigious journal Nature Photonics on November 14th, 2025, marking a significant milestone in the pursuit of faster, more efficient, and more scalable artificial intelligence. This breakthrough holds the promise of democratizing advanced AI capabilities, making them accessible and sustainable for a wider range of applications and industries, and fundamentally reshaping the landscape of computing and artificial intelligence for decades to come. The fusion of light and AI heralds a new era of computational power, driven by the elegance and efficiency of physics itself.