Top AI Networking Stocks to Watch, According to Rothschild By Investing.com

Rothschild highlighted four leading AI networking stocks—Lumentum, Coherent, Celestica and Credo—as beneficiaries of surging demand for optical components, Ethernet switching and AI infrastructure capacity. Lumentum cited ~80% additional capacity coming online over the next 18 months, Coherent and Lumentum both benefited from NVIDIA's $2 billion investment, and Celestica is seeing revenue growth above 50% with new 1.6Tbps AI switches available for order. Credo's acquisition of DustPhotonics and multiple analyst price-target hikes across the group reinforce the positive setup for the sector.

Analysis

The market is starting to price AI infrastructure as a multi-year capacity cycle rather than a one-quarter earnings story. The key second-order effect is that the winners are no longer just the hyperscalers buying chips, but the bottleneck suppliers in optics, photonics, and switching that can reprice faster as lead times tighten. That usually creates a stronger earnings revision impulse than the headline AI names because incremental capacity and mix improvement flow through with operating leverage.

The main opportunity is that the supply chain is still under-earning relative to demand. If hyperscaler capex remains elevated into next year, companies with near-term capacity adds and differentiated interconnect exposure can see a step-up in gross margin before the broader market fully notices, especially where product is moving from commodity transceiver economics toward higher-value integrated architectures. The risk is that this trade becomes crowded; the higher beta names can de-rate quickly if customers pause orders or if the market concludes the investment cycle is front-loaded.

A less obvious angle is competitive displacement inside the stack. As silicon photonics and co-packaged optics gain share, legacy cable and lower-spec interconnect providers face a structural headwind even if total AI spending stays strong. That means the market may be overestimating the durability of some revenue lines while underestimating the strategic value of companies that can participate in the architecture shift.

Near term, the setup remains constructive for months, not days: the catalyst sequence is order visibility, then margin expansion, then multiple expansion. The contrarian risk is that consensus is treating every capacity add as durable alpha, when in reality the best returns often go to the first movers and the enabling platform names, while late-cycle participants absorb price competition. The trade should emphasize balance-sheet quality and product differentiation over pure momentum.

Key Decisions for Investors

• Go long COHR vs. short a basket of lower-quality optical/component names over the next 3-6 months; thesis is that silicon-photonics and CPO exposure should sustain relative multiple expansion as architecture mix shifts.
• Buy LITE on pullbacks for a 6-12 month trade, but size smaller than COHR because the setup is more dependent on successful capacity monetization; take profits if the stock rerates ahead of visible margin inflection.
• Long CLS into the next two earnings prints, targeting the hyperscaler capex re-acceleration window over the next 2 quarters; pair with a short in a slower-growth networking hardware proxy to isolate AI switch share gains.
• Use call spreads in NVDA only as a hedge against the infrastructure basket rather than a primary long; the better risk/reward is in suppliers where capacity and pricing can surprise positively.
• Avoid chasing AEC-only exposure until post-order data confirms demand durability; if AEC displacement accelerates, that is a cleaner short than trying to fade the entire AI network complex.