Photonics Applications: Where the Technology Is Being Deployed

Photonics is not a single-use technology. In Jason Bodner’s Accelerated AI presentation, he identifies multiple applications where replacing copper with light creates value, and he frames the total addressable market across all of them at $40 trillion. That number aggregates several market projections, including the space economy ($1.8 trillion per McKinsey), humanoid robots ($5 trillion per Morgan Stanley), and the broader AI infrastructure buildout. It should be understood as a total addressable market over many years, not a specific investment return. But the breadth of applications is genuine and worth understanding.

Application 1: AI Data Centers

The primary application Bodner focuses on is inside AI data centers. Every major AI system, from ChatGPT to Meta’s recommendation engines to Google’s search AI, requires thousands of GPUs to work together. The data moves between those chips over copper wires, and copper is the bottleneck.

Photonics replaces copper with optical connections that use laser beams to transmit data at 124,000 miles per second. Broadcom debuted the industry’s first 400G optical chip for this application. Nvidia has invested over $7 billion in photonics companies to solve this problem for its own GPU clusters. For more on the core technology, see our photonics explainer.

One ChatGPT question burns 10 times more electricity than a Google search. A single AI data center eats as much power as 100,000 homes. Photonics reduces the power consumption dramatically because light barely uses energy and does not generate heat, which also reduces the cooling load.

Application 2: Satellite Networks

Elon Musk’s SpaceX is planning a fleet of one million AI-powered satellites. These satellites will need to communicate with each other and with ground stations at high bandwidth. Copper is not viable for satellite-to-satellite communication because of weight, power, and distance constraints. Photonics, specifically free-space optical communication, is the solution.

Bodner mentions this as a structural force driving photonics adoption. SpaceX is not the only company building satellite networks, but its scale makes it the most significant driver of demand for space-based photonic communication.

Application 3: Humanoid Robots

Morgan Stanley projects the humanoid robot market at $5 trillion. Humanoid robots require high-bandwidth, low-latency communication between their sensors, processors, and actuators. Copper wiring inside a robot adds weight, consumes power, and generates heat. Photonic connections solve all three problems.

This application is further out than the data center application, but it represents a significant long-term market for photonic technology. The companies that develop photonic components for AI data centers today will be positioned to serve the robotics market as it matures.

Application 4: Autonomous Vehicles

Autonomous vehicles are essentially mobile AI data centers. They process enormous amounts of sensor data in real time and need to communicate with each other and with infrastructure. The bandwidth and latency requirements are similar to those of AI data centers, and the power and heat constraints are even more severe because of the limited electrical capacity of a vehicle.

While Bodner does not focus extensively on autonomous vehicles in the presentation, the application is a natural extension of the photonics thesis. The same optical chips and transceivers that work inside AI data centers can be adapted for automotive use.

The Investment Implications

Bodner reveals Broadcom (AVGO) for free as the most direct play on the data center application. Broadcom’s AI revenue was $10.8 billion last quarter, up 143 percent year over year. The company designs custom AI silicon, builds networking switches, and makes the 400G optical chip. For more on Broadcom, see our Broadcom stock article.

The three bonus picks behind the paywall cover different segments of the photonics market. One is a pure photonics play that moves data as light between data centers, which is the inter-data-center application. Another builds parts that turn electrical signals to light inside AI clusters, which is the intra-data-center application. The third is a secret weapon inside every major AI data center, serving two of the three biggest AI spenders. For more on these picks, see our full Accelerated AI review.

The Acceleration Curve Context

Bodner frames these applications inside his Acceleration Curve framework. The internet went through a copper-to-fiber transition and produced returns of 16,000 to 188,000 percent for the infrastructure companies. Smartphones went through the same transition. Now AI, satellites, and robotics are going through it. The breadth of applications is what supports the $40 trillion total addressable market figure. For more on the pattern, see our copper to fiber article.

Considerations

The breadth of photonics applications is genuine, but each application has its own timeline and adoption curve. AI data centers are the most immediate and best-funded application. Satellites, robotics, and autonomous vehicles are further out and more speculative. Investors should focus on the applications with the most near-term revenue and use the longer-term applications as upside optionality.

If you want to explore the full thesis, you can access the Accelerated AI presentation through Brownstone Research.

This is not financial advice. Always do your own research before investing.