Understanding AI Enablers: A Fund Manager's Framework for Tech Investment Success

In the rapidly evolving landscape of artificial intelligence investment, the distinction between different categories of beneficiary companies has become crucial for portfolio construction. Denny Fish, portfolio manager at Janus Henderson, recently shared insights that illuminate how technology investors can navigate the complex ecosystem of AI-driven growth. His analysis centers on a powerful conceptual framework that helps investors understand not just which companies will benefit from AI, but how and when those benefits will materialize.

The enabler concept represents the foundational layer of this AI investment thesis—companies and infrastructure that make artificial intelligence deployment possible at scale.

The Three-Layer AI Investment Framework: Defining Enablers and Their Role

When discussing how to categorize AI opportunities across the market, Fish explains that the adoption curve unfolds in distinct phases, each populated by different types of companies. The framework he employs—which guides both the Janus Henderson Global Technology and Innovation Fund and the firm’s dedicated AI ETF (JHAI)—divides beneficiary companies into three primary buckets.

Enablers form the crucial foundation. These are the companies and infrastructure providers that make AI infrastructure possible. This category encompasses semiconductors, GPUs, ASICs, semiconductor foundries, equipment manufacturers, power producers, and data center operators. The enabler layer spans across technology, energy, and industrials—essentially all the hardware and infrastructure required to train large language models and then deploy those models for inference across various applications.

The reasoning is straightforward: before any software application or business process can be enhanced by AI, the underlying computational infrastructure must exist. Nvidia’s GPUs, TSMC’s manufacturing capacity, and companies providing power infrastructure all function as enablers. Without them, the entire AI ecosystem cannot function.

Beyond Infrastructure: Enhancers and End Users

While enablers capture headlines and drive short-term returns, the framework extends further to capture the full investment opportunity. Enhancers represent companies that possessed strong business fundamentals before AI arrived but will see those advantages compound through AI integration. Software companies with established market positions, critical data moats, and valuable customer relationships represent classic enhancers—they can embed AI into existing products to strengthen their value propositions. Consumer internet companies similarly function as enhancers, with AI poised to deepen user engagement and improve operational efficiency across both digital and physical dimensions.

End users represent the third layer—companies across healthcare, financial services, agriculture, and insurance that will deploy AI aggressively to reduce costs and drive revenue growth. Industry leaders in these sectors possess the scale and distribution infrastructure to capture significant competitive advantages from AI deployment, extending their market dominance through this technology shift.

This three-tier structure recognizes that AI benefits don’t materialize uniformly. The infrastructure enablers generate returns earliest and most dramatically. The enhancers see benefits accrue gradually as they integrate AI into existing products. End users experience the most profound but longest-delayed benefits as competitive AI adoption transforms their industries.

The Current Investment Cycle: Where Value Concentrates in 2026

The past three years have starkly illustrated this framework’s validity. Companies positioned as enablers—particularly semiconductor manufacturers and AI infrastructure providers—have delivered exceptional returns. The AI semiconductor ecosystem has benefited from fundamentals that exceed expectations, with earnings surging even as stock prices have risen, actually compressing multiples in many cases.

Software, by contrast, endured a difficult period. Despite revenue growth and earnings expansion in some categories, the entire sector struggled against perceived disruption threats and fundamental questions about AI’s revenue-generating potential within traditional software business models. The dispersion between winners and losers widened dramatically—companies that could convincingly demonstrate AI-driven value creation outperformed those without clear AI monetization pathways.

This pattern reveals an important truth: not all tech stocks participate equally in AI cycles. The practitioners who speak directly with industry participants, attend major conferences like the UBS technology conference, and conduct field research observe clearly that enablers continue to expand their capabilities. Nvidia’s announcement at CES regarding Vera Rubin, the next-generation GPU architecture, exemplifies this progression—systems that become simultaneously more powerful and more efficient, driving down the cost per token for model inference while maintaining performance gains.

Market Rotation and 2026 Outlook: Watching for Enabler Saturation

As 2026 unfolds, the investment landscape appears poised for continued dispersion among large-cap technology companies. Google’s impressive performance gains in early 2026 contrast with Meta’s softening momentum—a striking reversal from 2025 when the positions flipped. Such reversals suggest that even within the mega-cap tier, fundamental execution diverges.

The outlook for enablers remains constructive but increasingly nuanced. The foundation of enabling infrastructure will continue supporting the AI ecosystem, but investors should expect the category’s dominance to evolve. Software, having underperformed for three years relative to semiconductors and infrastructure, presents emerging opportunities as valuations compress and companies prove their ability to integrate AI meaningfully into operations.

The large-cap technology companies themselves will likely demonstrate bifurcation. Those extending competitive advantages through aggressive AI deployment—whether through superior chips, leading cloud platforms, or dominant market positions—should continue performing well. Others face pressure despite belonging to globally recognized brands.

Real-World Validation: From CES to Autonomous Vehicles

The confidence in this framework gains validation through real-world observation. Autonomous vehicle technology at CES 2026 illustrated the practical deployment of enabler innovations. Waymo’s San Francisco operations have matured to the point where experienced users actively prefer the service to traditional rideshare. Tesla’s Full Self-Driving system has advanced significantly but remains behind Waymo’s capabilities, reflecting the different technology pathways these companies pursue.

Wayve, a London-based autonomous vehicle company backed by SoftBank, Microsoft, and Nvidia, recently completed 45-minute autonomous journeys through London traffic without human intervention—a complex test case that demonstrates the rapid progression from theory to practical deployment. These real-world deployments validate the enabler thesis: without Nvidia’s GPU technology, without cloud infrastructure providers, without semiconductor foundries operating at full capacity, these demonstrations wouldn’t be possible.

The robotics and humanoid sectors emerging at CES similarly demonstrate how enabling infrastructure—chips, algorithms, training infrastructure—creates the foundation for entirely new product categories.

The Competitive Complexity: When Companies Blur Categorical Lines

The framework, while useful, becomes more nuanced when applied to hyperscalers. Microsoft exemplifies the blurred boundaries: Azure represents pure enablement infrastructure, while Copilot’s integration across Office and productivity applications positions Microsoft as an enhancer. Amazon similarly operates across multiple categories—AWS enables enterprise AI, while Amazon’s physical logistics infrastructure positions the company to capture tremendous efficiency gains from physical AI deployment through robotics and automation.

This complexity reflects a deeper reality: companies with both software and infrastructure capabilities can capture value across multiple phases of AI adoption. The hyperscalers’ competitive advantages—TSMC’s unparalleled manufacturing capabilities, Microsoft’s enterprise relationships, Amazon’s logistics network—create defensive moats that competitors struggle to replicate despite enormous capital expenditures.

Portfolio Construction: Balancing Resilience with Optionality

Implementing this framework within actual portfolio management requires balancing two competing objectives: resilience and optionality. The philosophy allocates 50-70% of the portfolio to companies offering resilience—businesses generating high returns with manageable outcome ranges, supported by strong competitive advantages and innovative leadership teams. These core positions represent holdings that could reasonably be maintained for five-year periods, assuming fundamental assumptions remain intact.

TSMC exemplifies this resilience category. Regardless of whether Broadcom, AMD, or Nvidia emerges as the leading semiconductor designer, “all roads flow through” Taiwan and the emerging Phoenix, Arizona manufacturing facilities. The foundry business model concentrates manufacturing risk in predictable ways—no matter which chip wins adoption, TSMC manufactures it.

The remaining 30-40% of the portfolio populates with smaller positions in companies with wider outcome ranges—tomorrow’s winners that might graduate into resilience positions over time. These positions accept higher uncertainty because they’re fundamentally bets on future competitive advantage development. The strategy acknowledges that future leadership remains partly unknowable; investment sizing reflects this reality.

Looking Forward: The Ongoing Evolution of AI Investment

As AI infrastructure deployment matures and the adoption curve progresses, the percentage of assets allocated to each framework category will shift. Companies that function currently as pure enablers may evolve into more balanced positions as their technology becomes commodity-like. Enhancers may transition into end-user categories as AI integration deepens. This dynamic approach to categorization, rather than treating it as static, explains why the framework structure itself—embodied in JHAI and similar actively-managed funds—adapts allocations over time.

The investment landscape heading into 2026 rewards practitioners who maintain direct contact with industry participants, verify assumptions through field research, and resist Wall Street noise in favor of fundamental data. The enabler infrastructure continues expanding its capabilities while driving down costs—a combination that ensures the AI infrastructure ecosystem remains healthy through the cycle. Yet dispersion between companies suggests that simple “buy tech” strategies will underperform. The distinction between enablers, enhancers, and end users increasingly determines investment success, validating the framework’s practical application for portfolio managers building resilient, adaptive positions in the AI age.