Lumentum: The Optics Play That Bridges Power and Memory

Reviewed 2026-09-06. Tape: 871.18 → 881.26 (+1.16%), worst weekly close −18.05%, peak-to-trough −26.45%, vs QQQ −16.49pp. The falsifier's first leg remains PARTLY FIRED (the co-packaged-optics timing moved outside the call's window on the 11 Aug guide); re-examined and it holds — the one apparent contradiction (a CEO quote on "no change in timing") is dated 16 Dec 2025, pre-dating the guide that moved. Coherent's 6 Sep dossier: CPO cannibalisation has NOT fired on the demand side (1.6T pull-ins accelerated, InP capacity binding), which bears on the second leg. The review proposes RESOLVE; Connor: verdict held for his own ruling in /admin/market-calls — the row stays ACTIVE with this logged.

Sources filled from the shelf and the 2026-09-06 review docs (catalogue item 3). 2 entries.

The quarter answered the question the 3 August entry asked. The stock gave the answer back within a week.

Lumentum reported FY2026 on 11 August. Full-year revenue was $3.014B against $1.645B the prior year — up 83% — verified from the 10-K filed 17 August (SEC accession 0001628280-26-057358). The fourth quarter came in around $1.01B, a record, with the guide raised. Nothing in the print was soft.

The 3 August entry, written when the stock was down roughly a third from its high, asked one question: was the sell-off sentiment or thesis? It argued sentiment — no guidance cut, no lost customer, no earnings miss — while stating plainly that "being right about a bottleneck and being right about when the market pays you for it are different skills." That entry cited $808.4M, the then-current quarter. It has been overtaken by this one; it is left standing unedited because a dated record that is corrected in place is not a record.

What happened next is the part worth writing down. The print landed on 11 August with the stock at $805.67. It rose to $938.01 the next session and peaked at $991.86 on 17 August, +13.8% against the 13 April call price. Then it fell 15.4% in two sessions — $991.86 to $839.49 on 19 August — closing at $857.70 on 20 August, which is 1.6% BELOW where the call was opened.

So: a record quarter, a raised guide, and four months after the call the position is flat-to-down and roughly back where it sat before the print. The 3 August question is now answered from both sides — the fundamentals were never the problem, and being right about them did not get paid. That is the distinction the earlier entry named, observed rather than predicted.

This is an observation, not a signal. Nothing here is a reason to add, trim or exit. The stated policy is that a position is exited on a broken thesis, not on price, and nothing in this print breaks the thesis — it is the strongest confirmation of it the call has had. What has changed is the evidence about timing, and that is worth having on the record before it is remembered more conveniently.

The widest gap in the book between what the company reported and what the stock did.

Down roughly 33% from a 52-week high of $1,085.68. What did not happen: no guidance cut, no lost customer, no earnings miss. The most recently reported quarter showed revenue of $808.4M, up about 90% year over year, with earnings per share ahead of consensus.

The decline is sector contagion plus sentiment. Coherent and Ciena sold off following Corning's print, and the optics complex went with them as traders re-examined the durability of AI capex generally.

This call was opened at HIGH conviction on the argument that optics is the bottleneck bridging power and memory. Nothing in the reported numbers contradicts that. What the drawdown demonstrates is that being right about a bottleneck and being right about when the market pays you for it are different skills. This is the position in the book where that distinction is being tested hardest, and it deserves an explicit decision rather than drift.