
Silicon Photonics: A New Approach to Acceleration
Harnessing Light for Faster, More Efficient FHE Computing.

The Rise of Silicon Photonics
For decades, the world of computing has relied on electronic circuits to process information. But as demand for faster, more efficient computing grows, we’ve hit a wall—power consumption is skyrocketing, and traditional processors struggle to keep up.
Enter silicon photonics—a revolutionary field that uses light instead of electricity to transmit and process information.
Silicon photonics emerged in the late 20th century as researchers sought ways to improve data transmission in telecommunications. Optical fibres had already transformed long-distance communication, but the challenge was integrating this technology into microchips. By the early 2000s, breakthroughs in fabrication techniques enabled light-based components to be built directly onto silicon chips—leading to rapid adoption in data centres, high-speed internet, and advanced sensing technologies.
Today, silicon photonics is a cornerstone of modern computing infrastructure. But at Optalysys, we’re taking it a step further—using light to perform computations, not just transmit data.
Why Light? The Power of Optical Computing
Traditional electronic processors face two major bottlenecks:
- Speed Limits – Electrons moving through wires create heat and resistance, limiting how fast processors can run
- High Energy Consumption – As computation increases, so does the power required, making large-scale processing unsustainable
Light, on the other hand, offers a game-changing alternative:
- Faster Processing – Light waves move at incredible speeds and can process information in parallel
- Energy Efficiency – Optical computing requires significantly less power than electronic circuits
These advantages make silicon photonics the perfect foundation for next-generation computing solutions—including Fully Homomorphic Encryption (FHE).


Optalysys: Unlocking the Potential of Photonic Computing
At Optalysys, we’ve developed a silicon photonics-based approach to computing that is specifically designed to accelerate FHE operations.
The diagram on the left illustrates the core principles of our approach:
- Laser light is injected into the system
- This light travels through a channel (called a “waveguide”) cut into the silicon
- The waveguide splits into multiple sub-channels, each carrying a portion of the original laser beam
- Modulators adjust the optical properties of each sub-beam, encoding information into the light
- All sub-beams are emitted into a free-space environment within the silicon chip
- As the beams travel, they interfere with each other, performing complex mathematical operations
- This interference naturally executes the core processing workload required for FHE computations
- The final output is detected and converted back into digital information
Real-World Applications
We have successfully implemented this technology in collaboration with the HEIR FHE compiler, demonstrating its potential for real-world applications. Our silicon photonics approach is paving the way for faster, more efficient, and scalable FHE computing.