The early days of artificial intelligence were defined by competition among large language models like OpenAI’s ChatGPT and Anthropic’s Claude. That phase, however, is largely old news.
Today, investor attention has shifted from who builds the smartest model to the massive scale of capital expenditures, or capex, behind the AI boom. This refers to money companies spend to acquire or upgrade long-term assets such as data centers, servers and networking infrastructure.
[Sign up for stock news with our Invested newsletter.]
Unlike regular operating expenses, these investments are meant to generate returns over several years. When a company pours money into capex instead of paying dividends or buying back shares, it’s a signal that management believes reinvesting internally will deliver stronger future growth.
Nowhere is this more evident than among the so-called “AI hyperscalers” — massive tech firms building and operating large-scale computing infrastructure to support artificial intelligence workloads. These companies are racing to dominate cloud-based AI services by spending heavily on energy-hungry data centers and specialized chips.
For example, a February research report from J.P. Morgan Asset Management projected that combined capex for four of the “Magnificent Seven” stocks — Microsoft Corp. (ticker: MSFT), Alphabet Inc. (GOOG, GOOGL), Amazon.com Inc. (AMZN) and Meta Platforms Inc. (META) — would reach $318 billion this year alone.
However, capital is largely circulating within a closed loop: hyperscalers buying chips from AI hardware makers, which in turn depend on those same hyperscalers as customers. Meanwhile, actual corporate spending on AI deployments and productivity gains remains uneven.
The question, then, is whether the industry is entering a speculative bubble — one that might painfully deflate once investors realize that the AI industry’s economics are more circular than revolutionary.
Bubbles are easy to recognize in hindsight, but during them, optimism and promises of a new paradigm attract continuous inflows and new highs. For retail investors, this enthusiasm has been most visible in the proliferation of AI-themed exchange-traded funds (ETFs).
For those not interested in picking individual AI stocks, these ETFs offer a hands-off, professionally managed way to invest in the artificial intelligence trend.
Some follow passive index benchmarks focused on AI-enabling technologies, while an increasing number are actively managed, supported by fundamental research and proprietary models aiming to identify the next big winners in the space.
“We believe it is critical to approach investing in generative AI companies with an actively managed approach,” says Thomas DiFazio, ETF strategist at Roundhill Investments. “The AI landscape is rapidly evolving, and it is crucial to be nimble.”
Here are six of the best AI ETFs to buy today:
| ETF | Expense ratio |
| Global X Artificial Intelligence & Technology ETF (AIQ) | 0.68% |
| Global X Robotics & Artificial Intelligence ETF (BOTZ) | 0.68% |
| Global X Data Center & Digital Infrastructure ETF (DTCR) | 0.50% |
| VistaShares Artificial Intelligence Supercycle ETF (AIS) | 0.75% |
| Roundhill Generative AI & Technology ETF (CHAT) | 0.75% |
| Xtrackers Artificial Intelligence and Big Data ETF (XAIX) | 0.35% |
Global X Artificial Intelligence & Technology ETF (AIQ)
“We’re still in the early stages of the AI cycle, and proper diversification is extremely important — be it across company stages or geographies — because it’s difficult to pick a winner or two this early,” says Tejas Dessai, director of thematic research at Global X ETFs. “With a thematic exchange-traded fund, you’re following an idea as opposed to a complex strategy.”
Global X ETFs offers one of the largest lineups of AI thematic ETFs. Its flagship product is AIQ, which holds $7 billion in assets under management (AUM) across 88 companies represented by the Indxx Artificial Intelligence & Big Data Index. This ETF is diversified globally, with numerous Asian companies represented in its top holdings. AIQ charges a 0.68% expense ratio.
Global X Robotics & Artificial Intelligence ETF (BOTZ)
“When you think about smartphones, laptops or even mobile applications, lower prices and cheaper development costs didn’t shrink the market but expanded it as innovation accelerated,” Dessai explains. “AI could follow the same trajectory, embedding itself into the physical world — from factories and drones to delivery vans and buildings.” This thesis is reflected in BOTZ, another AI ETF from Global X.
AIQ’s portfolio reflects a generalized approach to AI investing, whereas BOTZ has a deeper focus on applied automation. As a result, BOTZ’s portfolio tilts more toward innovative industrial and health care firms actively deploying AI and robotics solutions. It also has a significant international focus, with just over a quarter of its holdings based in Japan. BOTZ charges a 0.68% expense ratio.
Global X Data Center & Digital Infrastructure ETF (DTCR)
An often-overlooked angle in AI investing is to think like a landlord. Every AI model requires data centers packed with high-performance chips, and those facilities are typically leased from real estate owners rather than operated directly by the tech giants themselves. Investors can gain exposure to this side of the AI boom through DTCR, which charges a 0.5% expense ratio and pays a 1.3% 30-day SEC yield.
DTCR tracks the Solactive Data Center REITs & Digital Infrastructure Index. Top holdings include major data center real estate investment trusts (REITs) such as Equinix Inc. (EQIX) and Digital Realty Trust Inc. (DLR). The portfolio also holds telecommunications REITs that have modernized in recent years to support faster, denser data transmission for cloud computing, including American Tower Corp. (AMT) and Crown Castle Inc. (CCI).
[Read: 10 of the Best REITs to Buy for 2025]
VistaShares Artificial Intelligence Supercycle ETF (AIS)
“If you run an analysis between the S&P 500 and most AI ETFs, you will see high overlap,” explains Adam Patti, CEO of VistaShares ETFs. “Not so with AIS, which consists mainly of the ‘picks and shovels’ of AI infrastructure via companies that most investors may not own.” The ETF’s composition tilts more toward semiconductors and hardware companies, with a higher allocation to Taiwan and China.
“AIS leverages our proprietary ‘Bill of Materials’ methodology that was designed by AI luminaries on our investment committee,” Patti says. “We analyze the supply chain, layer in the actual costs to build AI data centers and semiconductors, and then dig into the company financials to determine their true AI business pipelines and related revenue.” AIS charges a 0.75% expense ratio.
Roundhill Generative AI & Technology ETF (CHAT)
“CHAT selects stocks using a proprietary methodology that combines a transcript score and sector score to evaluate companies’ relevance to generative AI, factoring in their revenue, profit and R&D investment in AI technologies,” explains Dave Mazza, CEO at Roundhill Investments. “Companies are then scored and selected based on their exposure to AI, market capitalization and liquidity.”
This actively managed methodology has produced a fairly concentrated portfolio of 42 stocks, with Nvidia Corp. (NVDA) sitting at the top with a 6.7% allocation. Other Magnificent Seven constituents within CHAT’s portfolio include Microsoft, Amazon and Meta Platforms. Over the trailing year, CHAT’s actively managed strategy has strongly outperformed the market and peers, with a 64.4% return.
Xtrackers Artificial Intelligence and Big Data ETF (XAIX)
Thematic ETFs often carry higher costs due to their use of specialized indexes or active management. However, value-minded investors can still find affordable options in this space. One relatively under-the-radar pick is XAIX, which manages about $86.4 million in assets. It tracks the Nasdaq Global Artificial Intelligence and Big Data Index for a modest 0.35% expense ratio.
The benchmark behind XAIX is notably sophisticated. It selects companies based on their involvement in AI-related subthemes such as deep learning, image recognition, natural language processing and speech recognition. Importantly, the index relies on patent activity to identify up-and-coming leaders in these fields, an approach that emphasizes innovation rather than backward-looking financial metrics.
More from U.S. News
Can AI Pick Stocks? 5 AI Investing Apps to Try
AI Alliances: Power Partnerships in the Tech Race
7 Best Tech ETFs to Buy in 2025
6 of the Best AI ETFs to Buy for 2025 originally appeared on usnews.com
Update 11/14/25: This story was published at an earlier date and has been updated with new information.
