Goldman Sachs Says That This Is the Next Big AI Infrastructure Megatrend With a 9X Growth Potential. 3 Stocks You Can Buy to Capitalize on This $154 Billion Opportunity

AI-driven demand is creating a major supply shortage in optical networking, with Goldman Sachs estimating the TAM could rise 9x to $154 billion in two years. Lumentum reported non-GAAP EPS up more than 4x year over year to $2.37, Ciena posted 290% EPS growth and raised full-year guidance, and Coherent's EPS rose 55% to $1.41 while management is already taking orders for 2028 and signing LTAs into 2030. The article argues pricing power and backlog growth should support continued strong earnings momentum across the group.

Analysis

The market is starting to price optical as a bottleneck rather than a commodity, which is a meaningful regime shift. That matters because once a component category moves from cyclic to constrained, pricing power and backlog visibility can outrun unit growth for several quarters, especially when customers are forced to dual-source and pre-commit capacity. The second-order winner is the broader capex stack: module suppliers, test-and-measurement vendors, and specialty materials names should see follow-through as hyperscalers push not just more ports, but denser, higher-speed architectures.

The key nuance is that the current shortage is likely to create margin expansion before it creates durable volume normalization. That sets up a classic “good news becomes better news” phase for the names already winning share, but it also raises the odds of eventual over-ordering once procurement teams panic about lead times. The first crack to watch is not demand exhaustion, but whether customers begin stretching qualification timelines or shifting mix toward lower-speed alternatives to bridge capacity gaps.

The contrarian risk is that consensus may be underestimating how quickly supply response arrives once pricing peaks. If incremental capacity in 800G/1.6T ramps faster than expected over the next 6-9 months, the market could front-run margin inflection and compress multiples even while revenue keeps growing. Another risk is customer concentration: a handful of hyperscalers can delay or re-phase orders, which would hit the group simultaneously rather than staggered, making this a high-beta trade to any AI capex pause.

Net-net, this is still a favorable setup, but the cleaner expression is to own the strongest execution plus backlog visibility and fade the least differentiated exposure. The asymmetry favors near-term upside in estimates, while the medium-term risk is not demand collapse but supply normalization and multiple compression after the market prices in several years of scarcity.

Key Decisions for Investors

• Long CIEN on a 3-6 month horizon: strongest backlog visibility makes it the cleanest way to express the shortage thesis; expect estimate revisions to continue as orders convert, with upside if guidance is raised again.
• Long LITE / short a higher-quality semicap proxy on any 5-10% pullback: use as a tactical momentum trade, but keep stops tight because the name is highly sensitive to any sign of pricing normalization.
• Pair trade: long COHR / short legacy networking exposure if available; COHR has the best leverage to capacity tightness plus longer-duration customer agreements, giving it the best multi-quarter revenue visibility.
• Buy downside protection on the basket via 3-6 month puts if the group gaps further: the main risk is a sharp de-rating if supply announcements or customer pauses emerge before the next earnings cycle.
• Monitor hyperscaler capex commentary and lead-time metrics over the next 1-2 quarters; if lead times start falling while prices remain elevated, take profits on 30-50% of longs because the trade will shift from scarcity to cyclical normalization.