The next revolution in computing may not come from faster processors, but from rethinking how information moves

The first articles in this series argued that semiconductors have become the foundation of the modern economy, that compute is organised as a hierarchy of control and that energy is emerging as its next constraint. Yet another limit is quietly approaching. It lies not in processors or power generation alone, but in the way information moves through computing systems themselves.

For more than half a century, progress in computing followed a remarkably consistent pattern. Engineers built smaller transistors, increased processor performance and packed ever more computing power into increasingly compact devices. The digital economy, the internet and artificial intelligence all emerged from that extraordinary trajectory. It was a revolution built on electrons.

Today, however, that revolution is approaching a different kind of limit. Not because engineers have run out of ideas, but because physics is becoming increasingly difficult to negotiate. The next chapter of computing may therefore depend on something entirely different.Not faster processors. Light.

Every Computing Revolution Encounters Its Own Limits

Technological revolutions rarely end because a technology fails. They end because it succeeds so completely that its own limitations eventually become impossible to ignore.

Steam engines transformed industry before electricity replaced them. Copper cables connected continents before fibre optics became indispensable. Mechanical switching gave way to semiconductor electronics. Each transition occurred when the previous architecture could no longer support growing complexity efficiently.

For decades, fibre optics transformed communication between cities, countries and eventually between data centres. Information travelled across oceans as pulses of light rather than electrical signals. The next transition is fundamentally different.

Integrated photonics brings those same optical principles inside computing itself — onto chips, between chiplets and across advanced packaging. Light is no longer simply connecting computers. It is beginning to connect the components inside them.

Artificial intelligence is creating precisely such a moment. The challenge is no longer simply performing calculations faster. It is enabling enormous numbers of processors to communicate continuously.

Artificial Intelligence Has Changed the Nature of Compute

Traditional computing focused primarily on individual processors. Modern AI operates very differently. Large language models are trained across thousands of processors working simultaneously. Every processor continuously exchanges information with countless others. Data no longer travels occasionally through the system. It flows continuously.

As AI models become larger, communication increasingly determines overall performance. Moving information has become almost as important as processing it. That changes the engineering challenge completely. Compute is becoming communication.

The next revolution in computing is no longer defined by how fast processors calculate, but by how efficiently they exchange information.

The Physical Limits of Electrons

For more than half a century, electrical signals have carried almost every piece of digital information. They remain extraordinarily effective, but they also obey physical laws that cannot simply be engineered away.

Every electrical signal generates resistance, produces heat and becomes increasingly susceptible to interference as bandwidth continues to rise. Engineers have repeatedly found ingenious ways to push those limits further, yet every new generation extracts a higher price in energy consumption, cooling requirements and overall system complexity.

The bottleneck is therefore gradually shifting away from processor performance towards communication efficiency. This is not a technological failure. It is the inevitable consequence of physics.

A Different Way to Move Information

This is where photonics enters the conversation. Rather than transporting information through electrical currents, integrated photonics transports information using light.

At first glance, that sounds like an incremental improvement. It is not. Light behaves fundamentally differently from electricity. Optical signals generate far less heat. They are largely immune to electromagnetic interference. They can carry enormous amounts of information simultaneously over extremely high-speed connections. Most importantly, they change how computing systems can be designed.

Photonics should therefore not be understood simply as replacing electrical wiring. It enables an entirely different communication architecture.

An Architectural Shift

Professor Martijn van Heck has argued for years that integrated photonics should not be understood as a faster replacement for electronics. Its true significance lies somewhere much deeper. It changes the architecture of computing itself.

As computing systems continue to scale, communication increasingly dominates both energy consumption and overall system complexity. Improving communication therefore improves the performance of the entire system.

The objective is not to replace electronics altogether. Processors will continue to rely on transistors. Memory will remain electronic. Control systems will continue to use electrical circuits. Instead, photonics complements electronics precisely where electrical communication begins to reach its practical limits.

The future is unlikely to belong exclusively to electrons or to photons. It will belong to systems that intelligently combine both.

Beyond Faster Chips

Public discussions about artificial intelligence often ask which company will build the fastest processor. That question increasingly overlooks the larger transformation taking place.

Performance no longer depends solely on processor speed. It depends on how efficiently processors communicate. The semiconductor industry is gradually becoming a communications industry. Communication is becoming an optical challenge.

The companies shaping the future of AI may therefore not simply build better processors. They may redesign the connections between them.

Infrastructure Becomes Architecture

This shift reaches far beyond engineering. It affects economics. It affects energy. It affects industrial policy. It affects geopolitics.

Countries investing in integrated photonics are not simply supporting another technology sector. They are investing in the next generation of digital infrastructure. The competition is no longer limited to designing faster chips. It increasingly concerns the architecture upon which future compute will operate.

The transition from electrons to photons may ultimately prove as significant as the transition from copper to fibre optics during the rise of the internet.

The Next Foundation

Artificial intelligence is exposing one of the oldest assumptions in computing: that information should primarily move through electricity.

For decades, that assumption powered the digital revolution. Tomorrow, it may become its principal constraint.

The future of computing will not abandon electrons. It will combine them with light, using each where it performs best.

The next revolution in computing will therefore not simply produce faster machines. It will redefine the architecture through which intelligence itself moves.

Series Note

Part of The Substrate Economy, a series exploring how semiconductors, compute, energy and photonics are reshaping the physical foundations of the global economy.

---

Credit

AI-generated illustration by Altair Media

Caption

Every computing revolution begins by overcoming the limits of the previous one. The transition from electrons to photons may become the next architectural shift in the evolution of intelligence.