• Nokia announced a modular coherent-optics portfolio spanning 1.6T, 2.4T and 3.2T designs, but said sampling is expected only in mid-2027 and general availability in the second half of 2027.
  • A separate multi-rail in-line amplifier and optical line system is scheduled for the second half of 2026. Up to 70% lower TCO, 40-fold amplifier density and 160 fibre pairs per rack are Nokia comparisons, not disclosed customer results.

A launch announcement with two different clocks

On 16 March 2026, ahead of OFC in Los Angeles, Nokia announced a suite of coherent optical designs and a multi-fibre in-line amplifier. The portfolio addresses data-centre interconnect, campus, metro, regional, long-haul, terrestrial and subsea uses rather than one generic ‘AI networking tool’. AI is the demand frame; the products are optical transport components and systems.

The dates matter. Nokia expects the new coherent family to begin sampling in mid-2027, with general availability beginning in the second half of 2027. The separate multi-rail optical line system is slated for the second half of 2026. An announcement or future sampling window is not evidence of production shipment, deployment or customer acceptance.

The building blocks and their intended jobs

Nokia says its method combines four new digital signal processors with optical front ends made in indium phosphide and silicon photonics. The goal is to assemble common blocks into application-specific packages instead of engineering an entirely separate optical engine for each reach, performance and efficiency profile.

The disclosed set includes a 1.6T-capable coherent pluggable for IP-over-DWDM data-centre interconnect; a 2.4T-capable pluggable for thin transponders across terrestrial and subsea networks; a low-power 3.2T-capable coherent-lite design for shorter campus and enterprise reaches; double-sided pluggables for CPO-, LPO- and NPO-based switches; full-band transponders; and 2.4T and 3.2T embedded transponders. ‘Capable’ describes a design target, not measured throughput on every route.

The amplifier tackles fibre count and hut space

For spans beyond 80 kilometres, multi-fibre expansion can be constrained by space in in-line amplifier huts. Nokia’s multi-rail design is intended to amplify more fibre pairs inside the existing footprint as operators scale by adding fibres rather than extracting indefinitely more capacity from one pair.

Nokia says the system can deliver a 40-fold density increase over its comparison point and support 160 fibre pairs in one rack. Those figures concern amplifier density and physical consolidation; they do not establish a 40-fold increase in end-to-end network capacity, traffic, service revenue or reliability. The public release provides no named deployment or independent field result.

Up to 70% lower TCO is a bounded vendor claim

Nokia attributes up to 70% lower total cost of ownership to lower cost, space and power across application-optimized configurations. Neither the announcement nor the independent reports reviewed here publish the baseline product, route, traffic load, power price, equipment count, operating period or complete calculation behind the maximum comparison.

Actual economics will depend on reach, fibre condition, spectral efficiency, chassis and rack requirements, interoperability, sparing, software, operations and the timing of commercial silicon and optics. A maximum design comparison must not be converted into a universal customer saving, a measured energy reduction or revenue already won.

OFC reporting adds context, not deployment proof

Data Center Dynamics independently listed the announced form factors and future availability dates. RCR Wireless News reported from Nokia’s OFC briefing that the combined Nokia and Infinera teams described four technology blocks and 13 coherent solutions, reinforcing that this is a product-development architecture spanning several applications.

The public record still lacks named coherent-family customers, purchase volumes, prices, qualification results, route tests, interoperability results, shipments or recognized revenue. Useful follow-up evidence will be samples, production specifications, operator trials and comparable field measurements—not a repetition of the AI demand narrative.

What to watch

  • Whether mid-2027 coherent samples arrive on schedule and which products reach general availability in the second half of 2027.
  • Production specifications for reach, wavelength, power, form factor, spectral efficiency and supported host systems.
  • Named trials or customers for each coherent design and the late-2026 multi-rail optical line system.
  • The baseline, configuration and time horizon behind the maximum 70% TCO comparison.
  • Field evidence for 160 fibre pairs per rack, operational density, reliability and maintenance burden.
  • Interoperability with IP-over-DWDM, CPO, LPO, NPO and third-party line-system environments.

Sources