Demand for the optical networking chips that connect AI data centres is running ahead of what Ciena can supply, CEO Gary Smith told the Wall Street Journal in an interview. Smith, who led the company through the 2001 telecoms crash, said the current AI-driven surge is different from that earlier boom.
The comments point to the pressure AI workloads are placing on the fibre and optical hardware that carry data between servers and between data centres. Optical networking is one of the less visible layers of the AI build-out, but it is a bottleneck that investors and equipment makers are watching closely.
Why optical networking matters for AI
Ciena supplies optical networking systems that telecom operators and cloud providers use to move large volumes of data over long distances. AI training and inference clusters require enormous amounts of bandwidth between chips, racks and sites, which pushes demand for exactly this kind of equipment.
Smith's comparison to the 2001 telecoms crash is notable because that period ended in a massive oversupply of fibre capacity. His argument, as reported, is that AI demand is outpacing supply rather than the other way around — a signal that the current cycle is being driven by real usage rather than speculative network building.
Ciena's position in the AI supply chain
The interview adds to a wider picture of AI infrastructure spending that stretches from chips to cables. Ciena is not a household name in consumer tech, but its order book is one way to gauge how fast data centre operators are expanding.
Techmeme flagged the Wall Street Journal interview as a key story, grouping it with other infrastructure and policy news. No specific financial figures or delivery timelines from Ciena were included in the source material.
