INDUSTRIAL REVOLUTION: AI buildout is far from finished, analyst insists

Key Concepts

• Optical Networking/Optical Interconnects: The critical hardware infrastructure required for AI data centers, utilizing specialized materials like Gallium Arsenide (GaAs) rather than standard silicon.
• Industrial Revolution Analogy: The perspective that the current AI build-out is a long-term infrastructure cycle (roads and railways) rather than a short-term "gold rush."
• Supply-Demand Imbalance: The phenomenon where demand for AI-related hardware is growing faster than the industry's ability to scale manufacturing capacity.
• Gallium Arsenide (GaAs): A specialized semiconductor material used in optical components, often a byproduct of zinc manufacturing, with supply chains heavily influenced by China.
• Market Rotation: The tendency for investors to rotate into "hot" sectors, often leading to volatility and investor fatigue between earnings reports.

---

1. The Long-Term AI Infrastructure Thesis

The analyst argues that the current AI market is not a fleeting "gold rush" but an Industrial Revolution-style build-out. While the market has been expanding for three and a half years, the analyst projects this cycle will last another five to ten years. The primary message to investors is to stop guessing when the cycle will end and instead recognize this as a "once-in-a-lifetime" opportunity to participate in the foundational infrastructure of the AI era.

2. The Criticality of Optical Technology

Optical components are essential for AI because they facilitate the high-speed data transfer required by modern AI clusters.

• Technical Distinction: Unlike standard consumer devices that use silicon chips (produced in billions via 12-inch wafers), optical components are significantly harder to manufacture.
• Material Constraints: The starting material, Gallium Arsenide (GaAs), is a byproduct of zinc manufacturing. Because China controls a significant portion of this supply chain, geopolitical factors (such as export restrictions on phosphate/related materials) create potential bottlenecks for global production.

3. Capacity Scaling and Pricing Dynamics

A common investor fear is that as manufacturing capacity increases, prices will collapse. The analyst refutes this by noting:

• Unprecedented Demand: Unlike previous cycles, demand is currently so far ahead of supply that even a projected ten-fold increase in laser capacity over five years is unlikely to lead to price erosion.
• Growth Metrics: The industry is moving from a capacity of 1,900 units to 22,750 units by 2027, yet the demand curve remains steeper than the supply curve.

4. Key Players and Market Positioning

The analyst highlights specific companies as primary beneficiaries of the AI infrastructure build-out:

• Lumentum (LITE): Identified as a market leader in the optical space. The analyst notes that Lumentum is capturing significant market share (growing from 2% to 6-7%) and is a "buy" despite high valuation expectations.
• Coherent (COHR): Described as a company catching up to Lumentum, also considered a strong "buy" with significant growth potential in the billions-of-dollars market.
• Fabrinet (FN): Highlighted as a manufacturing partner for major players like NVIDIA. The analyst praises the founder-led management team and views it as a high-quality, albeit volatile, way to gain exposure to the AI supply chain.
• Credo Technology (CRDO): While acknowledging the stock's parabolic rise, the analyst remains cautious. They note that Credo is not yet fully integrated into the "CXL" (Compute Express Link) ecosystem and lacks the gross margin profile or market position of Lumentum or Coherent.

5. Synthesis and Conclusion

The core takeaway is that the AI build-out is in its early-to-mid stages, characterized by a massive, sustained demand for specialized optical hardware. Investors are cautioned against "fatigue" caused by short-term quarterly volatility. The analyst emphasizes that the real value lies in the industrial-scale infrastructure—the "roads and railways" of the AI age—rather than speculative short-term trading. The most actionable insight is to focus on companies with established manufacturing capabilities and deep integration into the AI hardware stack, such as Lumentum and Coherent, while maintaining a disciplined, long-term perspective on the sector's growth trajectory through 2030.