An optics vendor backed by Nvidia plans to increase its wafer output by four times in order to keep up with surging demand for AI interconnects, according to The Register.
The details, as reported by The Register, are straightforward: the company intends to quadruple how many wafers it produces, and it is doing so specifically to meet demand tied to AI "interconnect" — the plumbing that shuttles data between the chips inside artificial-intelligence systems.
A quick translation for non-engineers: wafers are the thin discs of semiconductor material from which chips are cut. "Optics" here refers to components that move information using light rather than electrical signals over copper. As AI models grow, the systems that train and run them are built from large numbers of processors that must constantly talk to one another at very high speed. Optical interconnects are increasingly the way that data gets moved without becoming a bottleneck.
The Nvidia connection is notable. Nvidia is the dominant supplier of the chips used to train AI, and its backing of an optics supplier signals where it sees future constraints — not just in the processors themselves, but in the links that tie them together. A plan to grow wafer output four-fold is a sizable bet that this demand is durable rather than temporary.
The Register frames the move as a direct response to AI-driven demand, suggesting suppliers further down the chain are now racing to scale alongside the chipmakers grabbing headlines.
Why it matters: the AI boom is no longer only about who makes the fastest processors — it increasingly depends on whether the supporting supply chain, including the optical parts that connect those processors, can expand fast enough to feed it.