$LITE My own key takeaways from Lumentum’s fireside chat at the Mizuho Technology Conference is that the optical cycle still appears to be early in its architectural expansion.
This connects directly to the central argument in our recent connectivity article: connectivity is not one product category sitting next to compute. It is becoming part of the architecture that determines how efficiently compute can actually scale.
This is no longer just an 800G transceiver story. The opportunity is broadening across 1.6T, scale-up, NPO, CPO and optical circuit switching, while laser and InP capacity remain tight.
What stood out to me is that these architectures are not replacing one another. They are being added on top of an already strong base business. Scale-out keeps growing, scale-up pulls more optics inside and between racks, NPO may arrive before full CPO adoption, and OCS is developing into a separate opportunity around resiliency and traffic management.
That is exactly what we argued in the article (After Memory, I Am Watching Connectivity). As systems become denser, connectivity carries more of the burden. More bandwidth, lower latency, higher reliability and better power efficiency become architectural requirements rather than optional upgrades.
The more concrete bullish read-through for
$LITE is that management indicated new NPO opportunities are already absorbing added capacity. At the same time, the recent margin improvement has been driven mainly by pricing and product mix rather than cost cutting.
That matters because it suggests Lumentum is not simply participating in higher volumes. The company is starting to capture better economics as the bottleneck moves deeper into high-performance lasers, optical components and switching.
It also makes the case less dependent on one architecture winning. If NPO develops faster than CPO, Lumentum participates. If scale-up becomes more optical, Lumentum participates. If OCS grows into a much larger market, Lumentum participates there as well.
This is why I still read the cycle as early. Optical content per system is rising, the number of use cases is expanding, and much of the qualified capacity needed to support the transition is still being built.
I read the discussion as strengthening both the
$LITE thesis and the broader connectivity thesis. The architecture is becoming more optical, the economics are improving, and Lumentum appears positioned across several of the layers that will matter as the next generation of AI systems is built.
Another bullish nugget from
$LITE's CEO: The NPO opportunity "is frankly bigger than even the CPO opportunity".
Key takeaway: NPO or CPO, either way LITE wins.