Hyper Photonix has unveiled a next-generation 1.6T SiP optical transceiver product line. The company is positioning the 200G-per-lane Silicon Photonics (SiP) design for AI-driven data center networks that need higher bandwidth while keeping power and reliability in check.
Hyper Photonix plans to show “high performance 1.6T optical transceivers built on a 200G/lane architecture.” The company says Booth 449 demonstrations will cover 200G/lane operation “optimized for AI workloads,” along with “signal integrity and power efficiency,” a “scalable architecture designed for next-generation data center fabrics,” and “robust thermal performance under demanding operating conditions.”
For data center network designers, the key technical takeaway is the lane rate: 200G/lane is the building block for a 1.6T-class module, and it puts more pressure on end-to-end optical budgets, signal integrity, and module thermals than prior generations. Hyper Photonix’s focus on thermal performance and power efficiency is directly tied to what matters in dense AI clusters, where optics power draw and heat dissipation can become a constraint alongside switch ASIC power.
Volume production timeline and manufacturing expansion
Hyper Photonix said its 1.6T SiP transceiver product line will enter “volume production in Q2 2026,” supported by “new capital investment in highly automated manufacturing infrastructure.” The company also said it is building two new manufacturing facilities: a “second Asia-based high-volume factory” scheduled to begin operations in May 2026 in Hangzhou, China, and a “North American manufacturing site” in Monterrey, Mexico, opening in Q3 2026.
With those expansions, Hyper Photonix said its global manufacturing capacity will reach “more than 200,000 units per month in 2026,” and it linked that ramp to “accelerating demand from hyperscale data center customers,” referencing “recent commercial successes with leading hyperscale operators in Asia” without naming customers.
“The AI market is fundamentally reshaping data center interconnect requirements,” said Xavier Clairardin, CEO of Hyper Photonix. “Our 200G/lane 1.6T products deliver the performance, efficiency, and scalability required for next-generation AI clusters.”
Source: Hyper Photonix
