Logo Leicom ITEC AG
Spiegelung Gebäudefassaden Spiegelung Gebäudefassaden

The revolution of photonic processors for artificial intelligenc

21.01.2025

Despite the solutions currently proposed by Nvidia, it seems that the race for innovation in the chip industry is unstoppable: in addition to MIT, which has developed a revolutionary photonic processor, other research centers and companies are advancing cutting-edge studies in this field.

For instance, Stanford is investigating the integration of photonics and silicon for AI applications, while companies like Intel and IBM are developing photonic prototypes to accelerate data centers. In Europe, centers such as IMEC in Belgium are working on scalable technologies that combine electronics and photonics.

These collective efforts paint an increasingly competitive global picture: a photonic processor that uses light to perform complex calculations for artificial intelligence. This new technology could redefine the boundaries of computing power and energy efficiency, opening new horizons for the industry.

From Today's Chips to Photonic Chips: A Technological Journey

Companies like TSMC (Taiwan Semiconductor Manufacturing Company) represent excellence in transistor miniaturization, with manufacturing processes reaching 3 nm. These chips are at the heart of devices such as smartphones, computers, and data centers. On the other hand, China, with companies like SMIC, is working to reduce technological dependency, making significant advances in advanced chip manufacturing despite trade restrictions.

However, electronic chips face physical and energy limits: the more transistors integrated into a chip, the more heat is generated, and the more energy is required. This has driven research toward alternative solutions, including photonic chips.

Photonic chips, like the one developed by MIT, use light to transmit and process information. Light is faster than electrons and generates less heat, making this technology highly energy-efficient. Additionally, thanks to their modular architecture, photonic processors can perform operations in less than a nanosecond, overcoming the limits of traditional chips for intensive workloads.

These processors promise higher speed and the ability to process large amounts of data in real-time. The most promising applications include artificial intelligence, robotics and automation, telecommunications, and data centers.

Despite promising results, photonic processors are still in the development phase. One of the main challenges is mass production: although the technology is compatible with silicon manufacturing processes, integrating photonic and electronic systems on a single platform requires further improvements. Additionally, challenges related to precision and reliability in calculations must be addressed. Currently, the accuracy of MIT’s photonic chip exceeds 92%, but further optimizations are needed to achieve commercial applications.

Comparison with Quantum Computers

In a previous newsletter edition, we discussed quantum computers and their revolutionary potential for optimization and simulation problems. Compared to photonic chips, quantum computers operate in a different realm.

Photonic processors represent a technological breakthrough enabled by “photonic silicon,” which creates optical components directly on standardized silicon wafers. Quantum computers, on the other hand, require extreme operating conditions: just above absolute zero, with cryogenic cooling systems consuming kilowatts of energy. They also require isolation from vibrations and electromagnetic interference and take up entire rooms.

The applications reflect this fundamental difference: photonic chips excel in processing large amounts of data for deep learning and AI, with operations in less than a nanosecond and energy efficiency surpassing traditional chips. Quantum computers, however, specialize in non-deterministic problems, such as molecular simulations and advanced cryptography.

Despite the challenges – such as the current 92% accuracy of photonic chips and the stability of qubits in quantum computers – both technologies will coexist. Photonic processors offer a practical and scalable solution for the immediate future of AI and high-performance computing, while quantum computers represent a more distant but potentially revolutionary frontier for currently unsolvable problems.

These innovations redefine the boundaries of what is possible in the computing world and promise to transform industries such as robotics, telecommunications, data centers, and scientific research.

Next article of the newsletter, week 04:

Photonics in Cybersecurity: An Integrated Approach-->