These Nvidia-Backed Darlings are Great Dip-Buys, Say Pros

The article frames Nvidia’s AI “picks and shovels” opportunity as shifting from GPUs to photonics/optical connectivity, arguing Coherent, Lumentum, and Corning could benefit. It cites Street targets implying upside of ~44% for Coherent to $465, ~62% for Lumentum to $1,300, and ~42% for Corning to $270, suggesting analysts remain constructive despite recent pullbacks.

Analysis

The more important setup is not “AI keeps running,” but that the profit pool is migrating from raw compute to the bottlenecks that determine utilization. If hyperscalers keep pushing custom silicon, NVDA’s absolute dominance can persist while its relative upside compresses, because every incremental tensor accelerator still needs a fatter network and better optics to avoid stranded capex. That makes COHR and LITE the higher-beta expression of the second-order spend, while GLW is the slower, more diversified way to own the same theme with less operating leverage.

The market may be underestimating how quickly optical content can re-rate once the interconnect ceiling becomes visible in hyperscaler spending plans. The key catalyst path is the next 1-2 quarterly capex disclosures: if AI budgets stay intact, suppliers to transceivers, lasers, and fiber should see order visibility improve faster than headline semiconductor revenue. The risk is that this is a narrative trade before it is an earnings trade; if spending shifts from “build-out” to “optimization,” the multiple compression in the optics group can be severe because these names are priced for uninterrupted adoption.

Contrarian view: consensus may be too eager to declare optics the next obvious winner when the real winner could still be the platform owner capturing ecosystem control. If NVDA’s software, networking, and integrated rack strategy keeps customers locked in, the optics suppliers may remain important but not structurally scarce. Falsifiers are straightforward: a capex slowdown, any evidence that optical content per AI rack is flattening, or gross-margin disappointment in the next two reporting cycles.

Key Decisions for Investors

• Initiate a staggered long in COHR on weakness over the next 1-2 weeks; target a 20-30% upside over 3 months, with a thesis break if guidance or backlog fails to inflect on the next print.
• Prefer LITE as the cleanest relative winner in the optical chokepoint theme; consider a COHR/LITE basket for 3-6 months, but size modestly because both names are priced for near-perfect AI adoption.
• Use GLW as the lower-volatility AI infrastructure proxy rather than the high-conviction trade; suitable for a core holding if rates stay choppy, but expect less torque than COHR/LITE.
• For relative value, short NVDA against a long basket of COHR + LITE only if the objective is to isolate the supply-chain shift; otherwise avoid outright NVDA shorts because the stock still owns the base case of ongoing AI capex.
• Set an alert around the next hyperscaler capex commentary: if spending growth decelerates or optical content disclosures disappoint, de-risk the entire basket immediately.