For most of the last decade, the fiber optics industry ran on a comfortable two-year cadence. 100G gave way to 400G. 400G gave way to 800G. Each transition bought roughly two years of breathing room before AI infrastructure demands made the last generation look quaint. That cycle has now collapsed to something closer to twelve months, and the roadmap through 2029 is being written in real time on hyperscaler loading docks.

1.6T transceivers moved from sample quantities to real volume over the course of 2026. Coherent confirmed it was shipping 1.6T modules at OFC in March, and Nvidia's Vera Rubin platform, which entered volume production in the second half of this year, is the demand engine behind the ramp. Rubin pushes 200G per electrical lane, which makes 1.6T the natural optic for its Spectrum-X and Quantum-X fabrics. The hyperscalers absorbing those modules are not treating the milestone as a finish line. They are treating it as a waypoint. The question the industry is racing to answer: how do you move twice as much light through a connector that fits in your hand, and what do you do when you no longer can?

Members Only

Keep reading — it's free

Futures analysis is exclusive to FiberPulse readers. Drop your email to unlock every article.

No spam — unsubscribe anytime.