The Next Wave of AI Infrastructure: Why 'Pick-and-Shovel' Plays Offer Better Returns Than Core Hyperscalers

The AI revolution is no longer confined to the algorithms and software that capture headlines. As the technology matures, capital flows are shifting toward the foundational infrastructure that powers it-a trend creating opportunities for investors who look beyond the well-trodden paths of core hyperscalers. While companies like Meta, Google, and Microsoft dominate public discourse, the real value is being unlocked by the "pick-and-shovel" enablers: firms supplying data storage, semiconductors, power generation, and construction services. These underappreciated players are poised to outperform as the AI boom accelerates, driven by surging demand for energy, hardware, and physical infrastructure.

Data Storage: Sandisk's Resilience in a High-Demand Market

The exponential growth of AI models requires vast amounts of data to be stored, processed, and accessed rapidly. Sandisk, a leader in flash storage solutions, exemplifies how this sector is capitalizing on the AI infrastructure boom. In its fiscal fourth quarter of 2025, Sandisk reported revenue of $1.90 billion, a 12% sequential increase and above guidance, despite a GAAP loss of $23 million. The company's non-GAAP diluted earnings per share of $0.29 underscore its ability to navigate near-term challenges while maintaining a strong position in the AI-driven storage market. With guidance for Q1 2026 revenue of $2.10 billion to $2.20 billion, Sandisk's trajectory reflects the growing reliance on high-capacity, low-latency storage for AI workloads.

Semiconductors: ASML's Dominance in AI-Driven Chip Demand

At the heart of AI infrastructure lies the semiconductor industry, where ASML Holding NV remains indispensable. The Dutch company's Q3 2025 results highlight its critical role: total net sales of €7.5 billion, a gross margin of 51.6%, and net income of €2.1 billion. ASML's advanced lithography machines are essential for producing the high-performance chips used in AI accelerators and data centers. The company's Q4 2025 guidance-sales between €9.2 billion and €9.8 billion-signals continued strength, with full-year 2025 sales growth projected at 15% compared to 2024. As AI models grow in complexity, demand for ASML's cutting-edge equipment will only intensify, making it a cornerstone of the AI infrastructure ecosystem.

Connectivity and Construction: Amphenol's Record Performance

The physical infrastructure supporting AI-ranging from data center construction to high-speed connectivity-is another overlooked frontier. Amphenol, a global leader in electronic interconnect solutions, has seen explosive growth in 2025. Its second-quarter results included sales of $5.7 billion, a 57% increase in U.S. dollars and 41% organically year-over-year. GAAP diluted EPS of $0.86 and adjusted diluted EPS of $0.81 reflect the company's ability to capitalize on the AI-driven demand for robust connectivity solutions. Amphenol's success underscores the importance of construction and materials in enabling AI infrastructure, from fiber-optic cables to power distribution systems.

Energy Infrastructure: Tortoise Capital's Strategic Vision

Perhaps the most underappreciated enabler of the AI boom is energy infrastructure. Tortoise Capital, a firm specializing in energy and infrastructure investments, has positioned itself at the forefront of this shift. According to a report by Bloomberg, the firm is targeting companies like Williams Companies and Targa Resources, which provide behind-the-meter power generation and pipeline solutions for hyperscalers. For instance, Williams is already supplying power to Meta's data centers using existing infrastructure, enabling faster deployment and stable returns. Similarly, Targa Resources' Permian Basin operations offer a scalable energy solution for AI-driven data centers. Tortoise Capital's Tortoise AI Infrastructure ETF (TCAI), launched in August 2025, further underscores the firm's conviction in this space, focusing on electric power generation and data center infrastructure.

The energy demands of AI are staggering. According to data, data centers require continuous, high-capacity electricity, driving demand for natural gas and grid upgrades. Companies like GE Vernova, which supplies turbines for power generation, and Cipher Mining, pivoting from bitcoin to high-performance computing, are also gaining traction according to industry analysis. Tortoise Capital's strategy highlights a critical insight: while the AI narrative often centers on software and algorithms, the real long-term value lies in the energy and infrastructure assets that sustain it.

Why Pick-and-Shovel Plays Outperform

The case for pick-and-shovel plays over core hyperscalers rests on three pillars: contracted cash flows, lower volatility, and indispensability. Unlike hyperscalers, which face regulatory scrutiny and market saturation risks, infrastructure providers benefit from long-term, stable demand. Sandisk's storage solutions, ASML's chip manufacturing, Amphenol's connectivity, and Tortoise's energy assets are all essential to AI's growth, with limited substitutes. Moreover, these firms often operate in sectors with predictable margins and recurring revenue, offering a buffer against the cyclical swings that plague tech stocks.

Conclusion

As the AI revolution enters its next phase, investors must look beyond the familiar names to the enablers that make it possible. Sandisk, ASML, Amphenol, and Tortoise Capital represent a diversified portfolio of high-conviction plays, each addressing a critical gap in the AI infrastructure chain. With AI-driven demand for storage, semiconductors, connectivity, and energy set to surge, these underappreciated enablers offer a compelling path to outperformance in 2026 and beyond.