Janus Henderson’s 2026 survey of investors has identified some unusual dynamics when it comes to AI investment. Investors may be overlooking AI exposure in their portfolio, and seem indecisive on the volatility AI may introduce into markets in the short-term. The most obvious holdings, including chipmakers and mega-cap technology platforms, are only part of the story. Artificial intelligence is shaping capital spending, power demand, software business models and the economics of companies well outside the technology sector.
That makes AI exposure harder to measure than many investors assume. According to Janus Henderson’s 2026 Investor Survey, Perspectives on AI, nearly half of affluent and high-net-worth investors believe they have no exposure to AI. Given the ubiquitousness of the theme, it is virtually impossible for an investor to avoid the biggest secular theme of our lifetime. Janus Henderson Investors Wealth Strategist Ben Rizzuto points out that “this signals that advisors have the opportunity to educate their clients on what the AI investment is comprised of, and how much exposure they truly have.”
The survey also found that 58% of respondents would be uncomfortable having more than 10% of their portfolios allocated to AI-related investments. Before investors can decide whether that level is too high or too low, they need a clearer view of what they already own.
This reticence may be due to the interesting dynamic that 67% are concerned about an AI bubble or AI-driven market correction over the next 12 months, while 61% expect AI to have a positive impact on market returns over the next five years.
Those views can coexist. AI may be a powerful long-term force, even as parts of the market become vulnerable to short-term volatility if expectations move too far ahead of fundamentals. This dynamic is due to the well-known psychological effect referred to as Amara’s Law which suggests that we often overestimate what can be accomplished in the short-term and underestimate the long-term impact of transformative change.
A portfolio can hold no dedicated AI fund and will still have meaningful sensitivity to the AI theme. There are companies directly involved in AI development, and those who operate in industries that will benefit or be disrupted by AI, but few companies will remain entirely unaffected. The important question is whether that exposure is tied to companies with durable advantages or to businesses benefiting more from the AI narrative than from the economics of AI itself.
From our perspective AI is not a single investment opportunity, but rather three distinct opportunities in the enablers, creators, and beneficiaries of the technology.
See more: Rethinking the Trust Operating Model With AI in Mind
Agentic AI Is Expanding Demand Across the Infrastructure Stack
AI enablers provide the foundational elements AI requires, representing 'picks and shovels' investments. This category includes semiconductor companies manufacturing AI chips (especially GPUs), cloud computing providers offering scalable compute and storage, data infrastructure firms, and increasingly, power generation, transmission, cooling, and networking infrastructure essential for AI data centers.
Those areas remain important, and in parts of the ecosystem, demand for these components still exceeds supply.
While this investment opportunity is more mature than the others, it continues to evolve rapidly, changing the competitive landscape for companies ingrained in this space. We continue to see strong demand and persistent shortages across critical components, including advanced processors, memory chips, accelerators and optical networking equipment.
One reason the opportunity set is expanding is the rise of agentic AI. If we take a look at the earliest AI applications, they largely responded to individual prompts. Today, agents can reason, take action and coordinate multiple tasks with less direct human involvement. If businesses and consumers eventually deploy many agents that interact with each other, the volume of data created, stored, retrieved and transmitted could grow substantially.
That would increase demand for enabling technologies that were not always considered early AI winners. Central processing units, for example, can help schedule and execute agentic tasks. Memory chips are needed to store and retrieve the expanding amount of data these systems create, and optical components help connect increasingly powerful systems within and between data centers. Investors still need to judge pricing power, competition and capital intensity, but the AI supply chain is wider than it appeared two years ago.
The buildout also extends beyond traditional technology stocks. Power generation, grid capacity, cooling systems, construction services and industrial equipment are becoming more important as data centers expand. For investors with diversified portfolios, that means AI may also show up indirectly through companies financing or supporting the infrastructure behind the buildout.
From AI Spending to AI Value Creation
Recent earnings are starting to shed light on one of the central investor concerns: whether AI is showing up only in spending or also in growth. While the short-term investor focuses on capex resulting in some firms seeing free cash flow pressured or even turn negative because of massive AI spending, others have an eye toward the future ROI. That does not mean every dollar of AI capital spending will produce attractive returns, but it does suggest the cycle is not solely speculative.
Creators are businesses that are building AI, whether it is general use large language models or AI designed for a specific use case. Current economics suggest this may evolve into a winner-take-most environment, and the winners and losers are far from decided, making it an investment opportunity with a wide range of outcomes. These companies represent the next stage of the cycle.
The creators have introduced agentic AI which promises to enhance productivity and transform business models. Developers of innovative AI solutions have attracted many early adopters, with nearly 90% of companies using AI in at least one business function. Efficient integration of AI resulting in a distinct competitive advantage will become a hallmark of successful businesses in every industry.
Other business models face a more difficult path. AI is making many forms of routine cognitive work faster and less expensive, which can alter the source of competitive advantage. Some companies may see their advantages strengthened, while others may be weakened by the same force. Investors should examine whether AI improves a company’s competitive position, changes where value is created or simply adds a narrative that could take years to materialize.
Finally, end users form the broadest category and the final stage of the cycle. We are still in the early stages of AI adoption, but already companies and entire industries are being reshaped. Just as every company became a technology company to some degree over the past 25 years, all companies must decide how AI will enhance their business model, or face the consequences. However, given the speed of AI diffusion, companies don’t have 25 years to get on board.
Over time, the economic gains from AI should become more dispersed as adoption extends beyond the technology sector and the cycle evolves.
Audit What You Own
To assess their AI exposure, investors should start by considering whether their holdings are enablers, creators or beneficiaries. Portfolio names and category labels may understate how much AI sensitivity is already embedded in a portfolio. From there, they can determine whether their exposures align with their convictions, and how much exposure they are comfortable with.
Unless their money is under a mattress, their AI exposure will be greater than zero. The “right” amount of AI investment is situational, but investors should know what they own, take a long-term view (measured in years, not weeks) and determine the right amount of each of the AI investment types.
Our survey’s mix of near-term anxiety and long-term optimism captures the central investment tension. Investors may correctly anticipate AI’s transformative impact and still own the wrong company, the wrong layer of the value chain or the right business at an unjustifiable price.
AI is likely to create substantial wealth over time, but those gains will not be distributed evenly. Investors who understand the exposure already embedded in their portfolios will be better equipped to manage concentration, identify emerging beneficiaries and avoid mistaking a compelling theme for a uniformly attractive investment opportunity.
A message from Advisor Perspectives and VettaFi: Discover something new! Click here to register for our upcoming webcasts.
© Janus Henderson Investors
More ETF Topics >
