Jul 20, 2026 Equity
It’s hard to believe that it has only been around four years since the release of ChatGPT ignited a generative AI revolution. In that time, surging AI optimism has catalyzed massive capex in AI infrastructure, with over $5 trillion estimated to be spent by 20301 and remarkable advances beyond simply “text-in-text-out” to much more complex and productive capabilities. AI remains a powerful mega force still in the early innings of its adoption curve with the potential to reshape the economy on a scale comparable to the industrial revolution or rise of the internet. A range of companies are leading this transformation, hailing from various layers across the “AI tech stack” which includes infrastructure, intelligence, and agents.
Each layer of the tech stack plays an important role in AI’s emergence and presents different opportunities for investors. These layers don’t operate in isolation: infrastructure enables intelligence, while intelligence powers agents, and agents generate the real-world data that feds back into model improvement. As the AI ecosystem expands, opportunity may emerge across a wide range of companies working to power, train, secure, and deploy AI, including those tied to data centers, energy infrastructure, advanced computing, software, data, and automation. To delve deeper, we explore what is happening across each layer and how these areas may help shape the future of AI.
Source: BlackRock, as of July 2026.
Image description: Illustration showing the different layers of the AI tech stack across Infrastructure, Intelligence, and agents.
Infrastructure is a critical component of the AI tech stack, serving as the physical foundation for AI compute used in training large language models and running inference. Think, data centers, but also the power, cooling, networking, storage, and grid connections needed to keep them running.
This layer is becoming increasingly important as AI demand grows and power demand becomes a defining challenge. Worldwide AI infrastructure spending reached $318 billion in 2025, more than double the $153 billion recorded in 2024, and projected spending could reach $700 billion in 2026.2 Meanwhile, current estimates predict that power demand from AI data centers in the U.S. could grow more than 35x, with high-end industry estimates reaching 148 gigawatts by 2030, up from just 4 gigawatts in 2024.3
That means the infrastructure layer is no longer just about computing capacity. It is also about reliable power, cooling, grid connections, cybersecurity, and resilience. As AI becomes embedded in business processes, the systems powering it need to be dependable enough to support real-world workflows at scale. (Learn more about AI and infrastructure)
Imagine: You’re listening to the weather, and a record-breaking heatwave is predicted. Your first thought may be preparing for a potential power outage, but AI-powered grids could detect disruptions to power lines in real time, re-routing electricity without human intervention to spare you and your perishables.
The intelligence layer of the AI tech stack encompasses AI models and the unique data sets being used to train and improve them. If infrastructure is what powers AI, intelligence is what makes AI useful.
In the first phase of generative AI, much of the attention centered on model capability: how quickly models could write, code, summarize, search, or reason. The focus is now shifting towards business application: how AI can help companies improve productivity, reduce costs, create new products, and make better decisions.
Early evidence suggests AI is already creating measurable benefits in specific business functions. McKinsey found that while enterprise-wide profit impact remains limited, respondents most commonly report cost benefits from AI in software engineering, manufacturing, and IT. Respondents also report the greatest revenue benefits in marketing and sales, strategy and corporate finance, and product or service development.4
As AI adoption matures, the winners in this layer may not simply be those building the largest models. They may be the companies and platforms that can combine AI capability with proprietary data, distribution, workflow integration, and clear business outcomes.
Imagine: You can walk into a clinic and get a personalized cancer treatment plan within minutes — tailored precisely to your genetic profile and medical history. AI models can make this possible by turning mountains of clinical data into life-saving insights, transforming how doctors treat diseases.
The final layer of the AI tech stack is agents. If infrastructure powers AI, and intelligence turns data into insights, agents are the layer that helps turn those insights into action.
Think of agents as AI-powered systems that can reason, act, and execute tasks across digital and physical environments. Earlier AI tools were often used to summarize, draft, search, or answer questions. Agents are designed to go further. They can help coordinate workflows across software systems, customer service, research, compliance, supply chains, commerce, and other business processes. Interest in agents is already high, but adoption is still early — 62% of organizations in a survey found that they are at least experimenting with AI agents, while 23% report scaling agentic AI somewhere in the enterprise. However, no more than 10% of respondents report scaling agents in any individual business function.5
Agents are not just another AI feature; they may represent a new interface for software. Instead of users moving between multiple applications to complete a task, agents could help coordinate the steps across those applications.
Studies predict 33% of enterprise software applications to include agentic AI by 2028, up from less than 1% in 2024.6
Imagine: You can chat with an AI assistant about a birthday gift idea, and within seconds, it finds the perfect item, checks inventory, and completes your purchase — all without you lifting a finger. Ecommerce integration with AI chatbots like ChatGPT is turning casual conversations into seamless transactions.
As AI continues to evolve and change the way we view technology, so must the way in which we invest in it. The AI tech stack remains a useful framework. Infrastructure powers AI. Intelligence makes AI useful. Agents bring AI into workflows. Investing dynamically across the layers of the AI tech stack could be key to capturing the full investment opportunity in the AI theme.
While the most visible AI breakthroughs and largest companies in the space tend to capture headlines, we believe there are under-the-radar players in AI that represent some of the most compelling opportunities. An actively managed ETF like the iShares A.I. Innovation and Tech Active ETF (BAI), provides hand-picked exposure to companies enabling, developing, and deploying today's most advanced AI technologies.
Investors may get exposure to AI companies beyond mega-cap tech companies by investing in individual securities or funds that are unconstrained by market-cap. The iShares A.I. Innovation and Tech Active ETF (BAI), for example, invests in a concentrated portfolio of global AI and technology equities across all market capitalizations, chosen through bottom-up, research-driven fundamental investing.
BAI is the iShares A.I. Innovation and Tech Active ETF, an actively managed exchange-traded fund (ETF) that invests in a concentrated portfolio of global companies in the enablement, development, utilization and/or deployment of AI technology or products or services that leverage AI technology.
BAI invests in global AI and technology equities across all market capitalizations, chosen through bottom-up, research-driven fundamental investing.
AI-focused ETFs provide exposure to companies which may be positioned to benefit from long-term growth in artificial intelligence, across multiple industries. ETFs like BAI enable investors to seek diversified exposure to the AI theme without having to select individual stocks.
BAI provides exposure to companies across the global AI ecosystem, including:
Holdings may change any time as part of active portfolio management and market conditions. The fund holds approximately 50 securities, and allocations can be adjusted as investment views evolve.
BAI can be used as a thematic allocation within a diversified equity portfolio, complementing broader market exposure with targeted AI growth-potential.
The most current information — such as top 10 holdings — is available at iShares.com/BAI or in the official iShares BAI ETF fact sheet, which include holdings, performance, fees, and risk information.
Investors can invest through traditional brokerage platforms where they can also purchase stocks, bonds, and other ETFs.
As AI continues to evolve and change the way we view technology, so must the way in which we invest in it. Investing dynamically across the different layers of the AI tech stack could be key to capturing the full investment opportunity in the AI theme. While the largest names in the space tend to make headlines, we believe there are under-the-radar players in AI that represent some of the most compelling opportunities. An actively managed ETF like the iShares A.I. Innovation and Tech Active ETF (BAI), provides hand-picked exposure to companies enabling, developing, and deploying today's most advanced AI technologies.
Our Funds
ASSET CLASS
Investment Strategies
EXPLORE NOW
GET STARTED
GO FURTHER
FEATURED FUNDS
Insights
PORTFOLIO INSIGHTS
MARKET INSIGHTS
Education
EXPLORE NOW
INVESTING 101
ETF EDUCATION
INVESTMENT STRATEGIES & ASSET CLASSES
RETIREMENT SOLUTIONS
Resources
About Us
GET TO KNOW iSHARES
© 2026 BlackRock, Inc. All rights reserved.
1 Source: McKinsey. “The cost of compute: A $7 trillion race to scale data centers” April 28, 2025.
2 Source: IDC. “AI Infrastructure Spending Caps Historic Year at ~$90 billion in Q4 2025; 2029 Spending to Eclipse $1 Trillion.” April 16, 2025. Forward-looking estimates may not come to pass.
3 Data from PowerWork, as of March 2026. Forward-looking estimates may not come to pass.
4 Source: McKinsey. “The state of AI in 2025.” November 2025.
5 Source: McKinsey. “The state of AI in 2025.” November 2025.
6 Source: Gartner. “Gartner Predicts Over 40% of Agentic AI Projects Will Be Canceled by End of 2027.” June 2025. Forward-looking estimates may not come to pass.
Carefully consider the Funds' investment objectives, risk factors, and charges and expenses before investing. This and other information can be found in the Funds' prospectuses or, if available, the summary prospectuses, which may be obtained by visiting the iShares Fund and BlackRock Fund prospectus pages. Read the prospectus carefully before investing.
Investing involves risk, including possible loss of principal.
The Fund is actively managed and does not seek to replicate the performance of a specified index, may have higher portfolio turnover, and may charge higher fees than index funds due to increased trading and research expenses. There is no guarantee the fund will meet its investment objective.
Active funds are subject to management risk, which means the fund manager's techniques may not produce desired results, and the selected securities may not align with the fund's investment objective. Legislative, regulatory, or tax developments may also affect the fund manager's ability to achieve the investment objective.
Funds that concentrate investments in specific industries, sectors, markets or asset classes may underperform or be more volatile than other industries, sectors, markets or asset classes and the general securities market.
AI technology relies on large data sets, which can lead to inaccuracies. Companies in AI face competition, rapid obsolescence, and depend on demand from various industries. Regulatory scrutiny could limit AI development, with data collection facing closer examination and potential fines. Country-specific regulations could also impact AI and big data companies.
Technology companies may be subject to severe competition and product obsolescence.
International investing involves risks, including risks related to foreign currency, limited liquidity, less government regulation and the possibility of substantial volatility due to adverse political, economic or other developments. These risks often are heightened for investments in emerging/ developing markets or in concentrations of single countries.
This material is not intended to be relied upon as a forecast, research or investment advice, and is not a recommendation, offer or solicitation to buy or sell any securities or to adopt any investment strategy. The opinions expressed are as of the date indicated and may change as subsequent conditions vary. The information and opinions contained in this material are derived from proprietary and non-proprietary sources deemed by BlackRock to be reliable, are not necessarily all-inclusive and are not guaranteed as to accuracy. As such, no warranty of accuracy or reliability is given and no responsibility arising in any other way for errors and omissions (including responsibility to any person by reason of negligence) is accepted by BlackRock, its officers, employees or agents. This material may contain “forward-looking” information that is not purely historical in nature. Such information may include, among other things, projections and forecasts. There is no guarantee that any of these views will come to pass. Reliance upon information in this material is at the sole discretion of the viewer.
This material contains general information only and does not take into account an individual's financial circumstances. This information should not be relied upon as a primary basis for an investment decision. Rather, an assessment should be made as to whether the information is appropriate in individual circumstances and consideration should be given to talking to a financial professional before making an investment decision.
The Funds are distributed by BlackRock Investments, LLC (together with its affiliates, "BlackRock").
© 2026 BlackRock, Inc or its affiliates. All Rights Reserved. BLACKROCK, iSHARES, iBONDS, LIFEPATH, ALADDIN, THE MARKET IS YOURS and the iShares Core Graphic are trademarks of BlackRock, Inc. or its affiliates. All other trademarks are those of their respective owners.
GPS0726-5544287-EXP0727
