The AI boom is facing a physical-world challenge. As more and more data centres are built globally, the chips, cables, and components inside them are struggling to keep up with the volumes of data and speed of processing that ever-more capable AI systems require. Photonics—the use of light instead of electrons to transmit data—offers a way out. Startups in this space are developing optical interconnects, silicon photonics, and laser-based networking to bypass the bottlenecks of traditional copper wiring. These technologies promise faster, cooler, and more energy-efficient data transfer, critical for scaling AI infrastructure.
VCs are taking notice. Firms like Sequoia, General Catalyst, and Lightspeed have backed photonics startups in recent rounds. Companies such as Ayar Labs, Lightmatter, and Xscape are leading the charge, with products already in pilot or commercial deployment. Their solutions target everything from chip-to-chip communication to full-scale optical networking inside data centres.
The race is on to make photonics the default for AI-era computing. As models grow larger and data flows surge, the winners will be those who can make light the backbone of the digital nervous system.
Source: Sifted · Summarized by HeadlinesBriefing