What Is an AI ETF? Software, Infrastructure and Broad AI Funds Compared

By Adam Patti, Chief Executive Officer, VistaShares | November 2026

An AI ETF is an exchange-traded fund that invests in companies expected to be affected by artificial intelligence. The category is not uniform: some AI ETFs focus on software and applications, some on the physical infrastructure that AI runs on, such as chips, data centers and power, and some take a broad approach that spans the entire value chain. Two funds with “AI” in their names can hold very different portfolios, so understanding which type a fund is matters more than the label.

Why are there different types of AI ETFs?

Artificial intelligence is not a single industry. Building and using AI involves chip designers and manufacturers, memory and networking suppliers, cloud providers, data center developers, power producers and grid equipment makers, model developers and the software companies that build AI into their products. Our article on the AI supply chain describes these layers.

Because the theme spans so many industries, fund sponsors define it differently. Each definition leads to a different set of holdings, a different risk profile and a different sensitivity to how AI develops. Most AI ETFs fall into three broad groups.

What do AI software and application ETFs own?

These funds focus on companies that develop AI models or build AI into software products and services. Holdings may include enterprise software companies, cloud platforms, cybersecurity firms, data and analytics providers, and consumer internet companies.

The investment case for these funds rests on AI creating new revenue for software companies, through new products, higher pricing or broader adoption. Their results depend on whether software businesses can turn AI capabilities into sustained revenue and profits, and on how investors value that potential.

Software-focused AI funds often hold companies with high growth expectations and valuations. They can be sensitive to changes in interest rates and to shifts in sentiment about how quickly AI adoption translates into earnings.

What do AI infrastructure ETFs own?

These funds focus on the physical and industrial layers that AI depends on: semiconductors and semiconductor equipment, memory, networking and optical components, servers, data center developers and real estate, and the power generation and electrical equipment needed to run and cool AI computing. Our articles on the semiconductor supply chain and AI power and electrification cover two of these layers.

The investment case rests on spending to build AI capacity. As of November 2026, the four major U.S. cloud providers had guided to roughly $730 billion of combined capital spending for 2026, according to company guidance compiled in our report “The Deployment Gap.”¹ That spending flows to suppliers across the infrastructure stack.

Infrastructure funds can include businesses with very different characteristics, from chip designers to regulated utilities. Their results depend on the level and durability of capital spending, on supply constraints, and on the financing conditions that support large construction projects.

What do broad AI ETFs own?

Broad AI funds aim to capture the theme across the value chain, combining software, infrastructure and companies that use AI to improve their operations. They may hold large technology companies alongside smaller specialists.

Broad funds offer diversification across the theme’s layers, but they often hold many of the same large companies found in broad market indexes. That can reduce the distinct exposure they add to a portfolio that already holds those companies. Our article on how to evaluate a thematic ETF explains how to measure overlap.

How do AI ETFs choose their holdings?

AI ETFs use several methods to decide which companies qualify:

Revenue exposure. Some indexes require companies to earn a minimum share of revenue from AI-related products or services, or weight holdings by that share.

Classification and keyword screens. Some use industry classifications, patent filings or natural language analysis of company documents to identify AI-related businesses.

Expert or committee selection. Some indexes rely on a committee to judge which companies are relevant.

Active management. Actively managed AI ETFs rely on a portfolio manager’s research and judgment and can change holdings as the theme evolves.

The method matters. A revenue-based approach tends to favor companies whose results already depend on AI. A keyword or classification approach may capture a wider set of companies with smaller exposure. An active approach can adapt quickly but depends on the manager’s process. Our article on active versus passive ETFs explains the structural differences.

How should investors compare AI ETFs?

A practical comparison looks at several features side by side.

Holdings and layer mix. What share of the fund sits in software, semiconductors, data centers, power or other layers? This often explains differences in results more than any other factor.

Concentration. What share of assets is in the top ten holdings? Some AI funds are dominated by a handful of very large companies.

Overlap. How much of the fund is already held in the investor’s core equity holdings?

Methodology or process. How does the fund define AI, select companies and weight them? How often does it rebalance?

Geography. Does the fund include companies outside the United States, such as semiconductor manufacturers and equipment makers in Asia and Europe?

Costs and trading. What is the expense ratio, and how wide is the fund’s bid-ask spread? Our article on ETF bid-ask spreads covers trading costs.

What are the risks of AI ETFs?

Concentration. Many AI funds are concentrated in technology and in a limited number of large companies.

Valuation. Companies associated with AI may already reflect high expectations, and changes in sentiment can move prices sharply.

Spending cycles. Infrastructure funds depend on capital spending, which can slow if returns disappoint or financing becomes harder to obtain.

Technology change. Advances in efficiency, new architectures or new competitors can shift which companies benefit.

Regulation and trade. Export controls, tariffs and AI regulation can affect revenue and supply chains.

Investments involve risk, including possible loss of principal. More background on thematic investing is available in our thematic ETF guide.

Frequently asked questions

What is the difference between an AI software ETF and an AI infrastructure ETF?

An AI software ETF owns companies that build AI models or add AI to software products. An AI infrastructure ETF owns the companies that supply what AI runs on, such as chips, networking, data centers and power. They respond to different drivers: software funds to adoption and monetization, infrastructure funds to capital spending.

Do AI ETFs hold the same stocks as the S&P 500?

Many do. Large technology companies that are central to AI also make up a significant share of broad market indexes. The degree of overlap varies by fund. Comparing holdings by weight shows how much distinct exposure a particular AI ETF adds to an existing portfolio.

Are AI ETFs actively or passively managed?

Both exist. Many AI ETFs track rules-based indexes that define the theme through revenue, classification or committee selection. Others are actively managed, with a portfolio manager choosing holdings. Index funds publish their methodology, while active funds describe their process in the prospectus.

Why do two AI ETFs perform differently?

They often hold different parts of the AI value chain in different proportions. A fund concentrated in semiconductors can behave very differently from one concentrated in software or power, because each layer responds to different drivers. Differences in weighting, concentration, geography and fees also contribute.

How much of a portfolio should be in an AI ETF?

There is no single answer. Many investors treat thematic funds as satellite positions around a diversified core, sized according to conviction, risk tolerance and existing exposure to the same companies through broad market funds. A financial professional can help determine an appropriate allocation for individual circumstances.

This material is provided for educational purposes only. It is not investment, legal or tax advice, and it is not a recommendation or an offer to buy or sell any security. Information is drawn from sources believed to be reliable as of the date of publication, is not guaranteed as to accuracy or completeness, and may change without notice. Figures attributed to third parties are their estimates. Projections and forecasts are not guarantees of future results. Investors should consult their own financial, legal and tax professionals before making investment decisions. Investments involve risk, including possible loss of principal. Companies in a single industry or theme may be affected by common economic, regulatory or technological developments and may be more volatile than the broad market. References to companies, governments or programs are for illustration only.