Applied Optoelectronics Tumbles 9%, Coherent Sinks 8%, Lumentum Falls 4% as Optics Stocks Cool Off

AI optics names sold off on Friday, with Applied Optoelectronics down 9% to about $154, Coherent down 8% to around $346, and Lumentum down 4% to roughly $829, even though each remains sharply higher year to date. The move appears driven by profit-taking and sector rotation after recent AI infrastructure gains, not by company-specific negative news. Underlying fundamentals and guidance remain intact, including AAOI’s Q2 2026 revenue guide of $180M-$198M and Lumentum’s >85% YoY revenue growth outlook.

Analysis

The immediate takeaway is that this is a positioning reset, not a thesis break. In crowded AI infrastructure winners, the first derivative of fundamentals matters less than the second derivative of expectations; once a name has re-rated aggressively, even intact demand can’t prevent a multiple compression flush when marginal buyers step aside. The fact that the move is concentrated in optics while enterprise software catches a bid suggests a rotation out of the “most obvious AI beneficiaries” and into the laggards where earnings revisions have not yet been fully reflected in price.

That matters because optics names are now behaving like high-beta capex proxies: they will likely trade less on quarterly beats and more on hyperscaler commentary, channel checks, and delivery timing. The risk window is short-term, but the setup can stay unstable for weeks if the market starts questioning whether current order visibility simply pulled forward demand from 2027 into 2026. If that happens, the same stocks that looked supply-constrained can reprice sharply on even modest guidance hiccups.

The more interesting second-order effect is competitive. If AI optics momentum pauses, capital can rotate to adjacent beneficiaries with lower relative expectations, including enterprise software, networking, and even compute vendors that still look under-owned versus photonics. That does not mean optics leadership is over, but it does argue the easy money phase is over and the next leg higher will require fresh catalyst density rather than mere multiple expansion. In other words, the market is moving from narrative beta to execution scrutiny.

Contrarian view: the pullback may be useful because it resets positioning without damaging the fundamental setup. For long-only investors, that often improves forward returns in secular winners as forced de-risking clears weak hands. The wrong trade is to short the complex blindly; the right trade is to fade strength tactically or buy only on repeated proof that demand is still outrunning capacity, since the downside here is mostly valuation compression while upside requires another catalyst leg.

Key Decisions for Investors

• Trim/hedge AAOI, COHR, and LITE into strength over the next 1-3 sessions; treat them as crowded AI beta with 5-10% further downside possible if weekend risk-off spills into Monday open.
• For new exposure, wait for a 2-3 day consolidation before adding to LITE or COHR; preferred entries are on failed breakdowns being reclaimed, with a stop under the recent low and upside tied to next hyperscaler capex updates over 2-6 weeks.
• Consider a relative-value pair: long NOW / short a basket of AAOI+COHR+LITE for 1-4 weeks. The rotation signal is strong, and the pair isolates capital rotation from broad AI sentiment, with lower factor risk than outright sector shorts.
• If already long optics, buy short-dated downside protection rather than exiting fully: 1-2 month puts or put spreads on AAOI or LITE to protect against valuation air pockets while preserving upside if the AI thesis re-accelerates.
• Do not short the group outright into a pullback unless you have a catalyst calendar edge; the better risk/reward is to wait for a failed bounce, then express the view via a tactical short against a higher-quality AI hardware or software long.