Optical Supply Constraints Stretch to 2029: Planning AI Data Center Interconnects Early 2026-10-15
Publication date: 2026-10-15
According to Bloomberg, optical component capacity serving AI data centers has been booked through early 2029, and unmet demand for some products is still around 70% next year. Tight supply is lengthening lead times for high-speed interconnects. This article summarizes the report, explains what it means for AI server networking, and introduces our 1.6T OSFP 2×DR4 optical transceiver.
Why Optical Supply Is Tight Through 2029
On October 9, 2026, Lumentum CEO Michael Hurlston told Bloomberg in Tokyo that the company's optical component capacity for AI data centers is committed through early 2029.
The gap varies by product. For some products, roughly 70% of demand cannot be met through next year; for others, about 30% through 2028. Supply is tightest for key components such as indium phosphide (InP) lasers, which generate the optical signals used for interconnects inside AI data centers. Adding production lines takes several years, so the squeeze is not expected to ease soon.
What It Means for AI Data Center Planning
Optical links can become one of the items that set the pace of a build-out. Every server-to-switch connection needs a transceiver, and cabling and fiber capacity must be sized to match. When component lead times grow, a facility's go-live date can slip.
Teams planning new AI clusters can confirm transceiver specifications, fiber counts, and delivery schedules during the design phase and leave room for change. Early alignment between network, facilities, and procurement teams reduces the risk of late surprises.
1.6T OSFP 2×DR4: High-Speed Optical Links Between AI Servers and Switches
Our 1.6T OSFP 2×DR4 is a high-speed optical transceiver for data transmission between AI servers and network switches. It is designed with compatibility as a priority, and platform-level fit should be confirmed for each deployment. The specifications most relevant to cabling plans are listed below.
Data rate: 1.6T, providing high-bandwidth links between AI servers and switches
Form factor: OSFP, which requires switch or adapter ports that support OSFP
Interface type: 2×DR4, using two DR4 parallel-channel groups
Fiber count: 16 signal fibers per module (8 per MPO-12 connector), used to estimate fiber tray, patch cord, and patch panel capacity
Because each module uses 16 signal fibers, planners can estimate total cabling needs and confirm patching capacity before ordering.
Where It Fits and Where Limits Apply
1.6T links suit AI data centers that need more bandwidth per port between servers and switches. The actual benefit depends on switch port types, server network adapters, and overall network design, and delivery schedules should be confirmed project by project.
Planning Optical Connectivity as Part of AI Infrastructure
The report shows that AI infrastructure constraints extend beyond compute chips to the optical components that connect them. We recommend handling optical interconnect in early specification and scheduling discussions, which lowers scheduling risk later in a project.
Because new capacity takes years to arrive, enterprises planning their next AI deployment are better served by confirming optical interconnect options and delivery timelines early. We will continue to follow demand for high-speed optical interconnects and offer the 1.6T OSFP 2×DR4 for AI server-to-switch connectivity.
About Aeon Lighting Technology Inc.
We are a trusted global partner in AI infrastructure solutions, with deep expertise in data center architecture, high-density power distribution, intelligent cooling systems, and network cabling. Guided by rigorous quality control and compliance with international standards, we are committed to helping enterprises and partners worldwide accelerate digital transformation and advance toward the next generation of intelligent computing.
For more information, visit the Aeon Lighting Technology website: http://www.aeonlighting.com/