2026 is being called the first year of 1.6T by the industry. But if you are a decision-maker responsible for optical interconnect procurement for North American AI clusters, the truth you need to know is this: many vendors claim they “already have 1.6T mass production capability,” but only a highly concentrated few can actually secure large-scale orders from major North American customers.
This is not a question of vendor capability. It is a structural reality of the supply chain.
The first truth concerns the technology route. Silicon photonics has already become mainstream in the 1.6T generation, with a market share of 70%. With its advantages in low power consumption and low cost, silicon photonics has become a core direction for high-speed optical modules. Take leading industry players as an example: their in-house silicon photonic chips achieve a 95% yield rate and reduce costs by 30% compared with traditional solutions; their silicon photonic engines reach a 90% yield rate, far above the industry average of 75%. For buyers, the maturity of silicon photonics is now sufficient to support large-scale deployment decisions.
The second truth concerns supply chain barriers. Industry leaders have already locked up core materials through 2028. The supply gap for 200G EML lasers and silicon photonic components is genuinely constraining the pace of 1.6T volume ramp-up. Globally, only a small number of vendors have the ability to supply North American AI clusters at volume. Upstream materials—core chips such as 200G EML and high-speed DSP—have been locked up in advance by leading vendors through long-term orders. Even if second-tier vendors complete their module designs, they cannot produce at scale without enough core components.
The third truth concerns customer qualification. The full qualification cycle for major North American AI customers lasts 12 to 24 months. It is not only about hardware performance; it also includes stringent requirements such as long-term live network operation validation, factory audits, and localized overseas manufacturing. 1.6T high-speed packaging places extremely high demands on process control and signal integrity, and production yield ramp-up is difficult. This means that choosing a supplier that has already completed North American customer qualification and has a stable delivery record is far wiser than choosing a lower-priced but unqualified vendor.
For distributors evaluating 1.6T suppliers, it is recommended to focus on three dimensions: the status of long-term locking for core materials, live network deployment cases with North American customers, and the capacity ramp-up progress of overseas factories. These three indicators reflect a supplier’s real delivery capability far better than the parameters on a product datasheet.
