Lumentum is shipping 30% below customer orders for photonics components that power AI network infrastructure, signaling a sustained supply crunch that extends through calendar 2027. All externally modulated laser (EML) production capacity is committed under long-term agreements running through at least end-2027, leaving hyperscalers competing for limited allocation.
The company's optical communications (OCS) backlog has climbed past $400 million, with most units scheduled for second-half delivery. Management added indium phosphide manufacturing capacity by over 20% in the December quarter alone, with further expansions planned. Despite these additions, the demand-supply imbalance is widening rather than closing.
Lumentum now positions itself as a foundational supplier to virtually every major AI network deployment. The component shortage creates multi-year revenue visibility for specialized photonics manufacturers, contrasting sharply with typical tech hardware cycles that see rapid supply-demand rebalancing.
For investors, this presents a rare scenario where capacity constraints drive sustained pricing power and order visibility. Companies holding long-term supply agreements gain competitive advantages in AI infrastructure buildouts, while those without allocations face deployment delays or higher spot-market pricing.
The bottleneck sits at indium phosphide wafer fabrication, a capital-intensive process requiring 18-24 month lead times for meaningful capacity additions. This creates a structural moat for incumbent suppliers with existing fab infrastructure. Coherent, II-VI, and other photonics players face similar demand dynamics.
Supply chain risk materializes in two forms: hyperscalers may need to adjust datacenter deployment schedules based on component availability, and equipment manufacturers lacking photonics allocations could lose market share to better-supplied competitors. The 30% undershipping rate suggests current AI infrastructure plans exceed realistic near-term delivery capabilities.
Long-term agreements through 2027 lock in customer commitments but also limit supplier flexibility to serve new entrants or faster-growing accounts. This favors established AI players with early supply contracts over later-stage entrants attempting to scale infrastructure rapidly.