Laser Processing Units could give traditional CPUs, GPUs and quantum computers a run for their money – but don’t expect them to run Windows anytime soon
- LightSolver unveils Laser Processor Unit designed to solve physics equations at high speed
- Laser computing system uses optical memory to avoid processor and memory data bottlenecks
- Company begins offering early lab access with roadmap to one million variables by 2029
Laser-based computing is emerging as a potential challenger to established high-performance computing systems, with LightSolver revealing what it calls a Laser Processing Unit (LPU).
The Tel Aviv-based startup says its hardware can directly solve partial differential equations, a fundamental class of problems in physics and engineering.
Unlike quantum systems or GPU clusters, the approach is based on a grid of lasers acting in unison.
Mapping equations onto the LPU
LightSolver says its hardware can map equations such as the heat equation (which models heat flow) and Schrödinger’s equation (a core equation in quantum mechanics) directly onto the LPU.
By using the natural properties of lasers as electromagnetic waves, LightSolver’s tech avoids many of the constraints found in digital systems.
This includes bottlenecks caused by data transfer between memory and processor.
According to LightSolver, the LPU operates with embedded optical memory. In practice, this means laser states are preserved inside the resonator, allowing each calculation step to build on the last without needing to move data externally.
Sign up to the TechRadar Pro newsletter to get all the top news, opinion, features and guidance your business needs to succeed!
The result is constant-time iteration measured in nanoseconds, independent of problem size.
While LightSolver compares its platform to both high-performance and quantum computers, it stresses that it is not built on photonic chips.
Today’s photonic processors are two dimensional, whereas its laser-based design is three dimensional, which the startup argues will allow greater scalability.
LightSolver has set a roadmap to reach 100,000 variables by 2027 and one million by 2029.
The company is already beginning to make its hardware available for testing. An Alpha version of its processor and a digital emulator are being offered to researchers through what it calls its LPU Lab.
LightSolver believes this step will help scientists and engineers experiment with the technology before it reaches commercial grade.
“Classical computers digitize analog nature, and we pay the price in longer run times and wasted energy. By running large-scale physics simulations on a physical machine, we can solve them more efficiently than any HPC or quantum system available today,” said Ruti Ben-Shlomi, CEO and co-founder of LightSolver.
Research papers have been published on the subject, and presentations have been given at computing conferences, including ACM Computing Frontiers 2025.
The company has also formed partnerships with simulation software providers and is working with HPC centers and national labs.

You might also like
LightSolver unveils Laser Processor Unit designed to solve physics equations at high speed Laser computing system uses optical memory to avoid processor and memory data bottlenecks Company begins offering early lab access with roadmap to one million variables by 2029 Laser-based computing is emerging as a potential challenger to established…
Recent Posts
- NordVPN Coupons and Deals: 77% Off in June 2026
- You don’t need to spend a fortune on good audio — these 20 headphones under AU$100 have hundreds of 5-star user reviews
- Nintendo confirms it will sell a new Switch 2 with replaceable battery in the EU
- Apple begins requiring age verification for App Store use in Texas
- The co-creator of Scavengers Reign is working on a new show for Netflix
Archives
- June 2026
- May 2026
- April 2026
- March 2026
- February 2026
- January 2026
- December 2025
- November 2025
- October 2025
- September 2025
- August 2025
- July 2025
- June 2025
- May 2025
- April 2025
- March 2025
- February 2025
- January 2025
- December 2024
- November 2024
- October 2024
- September 2024
- August 2024
- July 2024
- June 2024
- May 2024
- April 2024
- March 2024
- February 2024
- January 2024
- December 2023
- November 2023
- October 2023
- September 2023
- August 2023
- July 2023
- June 2023