Computers that use light rather than electricity to represent and manipulate data could slash the power demands of data centres and simultaneously speed up calculations. Two studies published today describe breakthroughs in running real-world problems on light-based computers, making the technology on the verge of commercial application, say researchers.
Electronic computers, similar to those widely used nowadays, have traditionally adhered to Moore's Law, which states that their processing power would double approximately every twenty-four months. However, in recent times, advancements have decelerated due to the inherent physical limitations encountered during the downsizing of transistors.
Read more
How PsiQuantum aims to construct the globe's biggest quantum computer by 2027
Scientists are exploring various possible answers, like quantum computing and photonic computing. However, although quantum computing continues to face challenges in becoming truly practical, photonic computing has advanced to the stage where chip designs, exemplified in two recent research papers, are conducting real computations. Moreover, the very facilities that produce silicon chips for traditional electronics can also be employed to make these photonic chips.
Optical computers have significant potential benefits compared to their electronic counterparts. Firstly, since light particles travel quicker than electric charges within circuits, this might accelerate computations and minimize delays between computational steps. Additionally, due to the lack of resistance encountered by photons and their low absorption rate by chip materials, these devices could perform tasks with reduced power consumption relative to electronic systems, which often necessitate substantial amounts of energy for cooling purposes.
A Singapore-headquartered firm named Lightelligence demonstrates through their research that their apparatus, referred to as a photonic arithmetic computing engine (PACE), integrating both a photonic chip and a microelectronic chip, can effectively solve Ising problems. These types of challenges hold practical implications for sectors like logistics along with numerous others.
In the meantime, the U.S. startup Lightmatter asserts that its chip called Envise can execute the AI model BERT to generate text in Shakespeare’s style, achieving an accuracy level comparable to traditional electronic processors. Attempts to reach Lightmatter for confirmation were unsuccessful.
Bo Peng At Lightelligence, they mention that the sector is becoming crowded with startups and that the technology is quickly advancing. "We are essentially in pre-production," states Peng. "This is moving towards being an actual product rather than merely a laboratory showcase."
As the realm of quantum computers aims to showcase quantum supremacy—the juncture at which these machines deliver capabilities unattainable by traditional systems—Peng states his concentration lies on achieving photonic supremacy. While declining to specify a timeline, he indicates that the tech is nearing readiness for practical use cases, potentially through incorporating photonic chips into existing setups alongside conventional electronic components to enhance particular functions.
Read more
Google's advancement clears the path for big quantum computer development.
Notably, the hardware used by Lightelligence for their research was manufactured in a PCI Express format This represents the typical expansion card layout used in desktop PCs, enabling additional components like graphic cards to be installed. Their product can easily be integrated into various off-the-shelf desktop systems—though it would require appropriate software to interface properly with them.
Robert Hadfield At the University of Glasgow, UK, remarks that both studies indicate "we're approaching a critical juncture" in this field. He adds, "'This could soon reach a stage where businesses might view photonic processors as a practical option.' It’s fascinating to observe how far this design has evolved. Furthermore, these photonic chips were manufactured at one of the premier fabrication facilities globally, suggesting they have potential for large-scale manufacturing.”
Stephen Sweeney , also at the University of Glasgow, says that we have already seen optical data transmission rolled out around the world with fibre optics, and that photonic computing is now also close. "Photonics allows you to do things with higher speed and lower loss than you can do with electronics," says Sweeney. "And if you need to be able to do vast amounts of computation, you need to start looking at that."
Journal reference
Nature DOI: 10.1038/s41586-025-08786-6
Journal reference
Nature DOI: 10.1038/s41586-025-08854-x