A relatively quiet day for Serenity, the stock-picker tracked on this site, but the one item he flagged spoke directly to the thesis he has been pushing for months: the idea that the AI buildout's real scarcity lies not in the flashy chip names but in the humble light sources that feed optical networking.

Serenity pointed to fresh industry commentary from Foxconn — the Taiwanese electronics giant best known as Apple's contract manufacturer, and increasingly a builder of AI server infrastructure — which is reportedly weighing joint investments to head off shortages of upstream photonics materials. More pointedly, the chairman of its interconnect arm singled out CW lasers and optical fiber as the single biggest bottleneck in photonics.

That matters to Serenity because CW (continuous-wave) lasers are the light source that powers silicon photonics and co-packaged optics, or CPO — the approach of putting the optical engine right next to the switch or GPU die to raise bandwidth and cut power. His framework is built on "buying what's upstream of NVIDIA": he treats NVDA's roadmap — including its multibillion-dollar push into optical interconnect — as a signal of where demand will pile up, then hunts the smallest merchant suppliers of the components that everyone quietly depends on. A "chokepoint," in his language, is exactly this: a critical upstream part with few substitutes. Notably, he does not own NVDA itself, viewing it as a crowded mega-cap whose spending simply makes the upstream parts scarce.

He drew a parallel between the Foxconn signal and the dynamic already visible at two of his large-cap photonics holdings. Coherent (COHR) is the vertically integrated optical player he describes as doing "everything" — substrates, EML lasers, transceivers and silicon carbide — and which he holds as a steadier, lower-beta core position sized around 2.5% of his book, with capacity he believes is effectively sold out through 2027-28. Lumentum (LITE), his roughly 5% photonics weight, is the quality incumbent he ties to Google's TPU optics and broad hyperscaler capex. Serenity has previously highlighted a specific vulnerability at Lumentum: it has locked up its own EML (electro-absorption modulated) laser supply for high-speed links but is forced to buy CW lasers from rivals — precisely the shortage Foxconn is now naming.

The takeaway Serenity emphasized was one of validation rather than any new trade. He noted with some satisfaction that outside industry voices are now converging on the same CW laser bottleneck he has been arguing about publicly, a reiteration he sees as confirmation that the constraint is structural rather than a passing supply hiccup.

For readers new to the thesis, the significance is directional: Serenity generally uses COHR and LITE as the safer, benchmark large-caps against which he pitches his higher-upside smaller names further upstream in lasers, substrates and epiwafers. When a manufacturer the size of Foxconn publicly flags the same shortage he has staked his portfolio on, it strengthens the case for the whole chain — even if today brought no change to his positions. It was, in short, a day of confirmation, not action.

*This is derived commentary for information only and is not investment advice.*