Key Concepts
- AI Infrastructure Build-out: The shift in investment focus from semiconductors to the physical infrastructure (power, cooling, storage) required to support AI data centers.
- Liquid Cooling: A thermal management technology essential for high-density AI chips (e.g., Vera Rubin generation) that are too hot for traditional air cooling.
- Power Bottleneck: The supply-demand gap in electricity generation, with data centers driving a massive increase in demand.
- Energy Storage Systems (ESS): Technologies required to manage and stabilize power supply for data centers.
- Labor Availability Gap: A critical industry headwind involving a shortage of skilled labor required to construct and maintain complex data center infrastructure.
1. The Industrial Sector and the AI Narrative
Evelyn Cho (Neuberger Berman) argues that the AI investment story has evolved beyond mega-cap tech and semiconductors into the "infrastructure story." Industrial companies—specifically those in electrical equipment—are currently among the best performers in global indices. This growth is driven by a combination of cyclical uplift and the physical requirements of AI data centers.
2. Key Investment Picks
Cho highlights three companies addressing specific constraints in the AI ecosystem:
- nVent Electric (NVT): Solving Heat
- Role: A leader in liquid cooling technology.
- Technical Detail: Liquid cooling is currently used in 15–30% of AI deployments, a figure that increases significantly for cluster deployments. It is more efficient than air cooling for high-density chips like the Vera Rubin generation.
- Financials: Targeting 20%+ earnings compounding with mid-teens organic growth.
- GE Vernova (GEV): Solving Power
- Role: A major player in power generation, specifically gas turbines.
- Market Position: Holds >25% of the global gas turbine market share.
- Financials: Boasts a $200 billion backlog through 2027, providing high visibility into growth. The company benefits from high Average Selling Prices (ASP) and a robust, high-margin services engine.
- Context: Cho views the recent stock pullback as a buying opportunity, noting that while peers plan capacity additions for 2029, GEV is positioned to meet immediate demand.
- Contemporary Amperex Technology Co. Limited (CATL): Solving Energy Storage
- Role: A leader in integrated full-stack AI data center energy storage systems.
- Innovation: Pioneering sodium-ion batteries as a medium-term storage solution.
- Global Reach: Demonstrates that the AI opportunity is not limited to North America; the company is successfully accessing the U.S. market through partnerships (e.g., Ford).
3. The Power Scarcity Theme
Cho identifies power scarcity as a defining investment theme for the next decade.
- Data: Transformer investment has increased by 120% over the last five years, yet capacity remains insufficient.
- Demand Shift: AI data centers are fundamentally reshaping the energy landscape, causing demand to grow after years of stagnation.
- Requirement: Estimates suggest a need for 15+ gigawatts of increased power generation annually to keep pace with data center growth.
4. Industry Headwinds: The Labor Constraint
Despite the technological advancements, Cho identifies labor availability as a primary risk.
- The Problem: Engineering and Construction (E&C) companies report a significant shortage of the highly skilled labor required to build data centers.
- Impact: This "gating factor" could limit the speed at which power and cooling infrastructure can be deployed, potentially slowing the overall AI build-out toward the end of the decade.
5. Synthesis and Conclusion
The AI build-out is transitioning from a software and chip-centric narrative to a physical infrastructure-intensive phase. Investors are increasingly looking toward "brick-and-mortar" industrial players that solve the three primary bottlenecks of AI: heat (nVent), power (GE Vernova), and storage (CATL). While these companies offer strong growth profiles and high visibility, the sector faces a significant structural challenge in the form of a skilled labor shortage, which may act as a ceiling on the pace of infrastructure expansion.
AI summaries can miss context or contain errors. Check important details against the original video.