The AI Build-Out Is Just Getting Started

In this episode of Motley Fool Hidden Gems Investing, Motley Fool contributor Rachel Warren talks with Jay Jacobs, U.S. Head of Equity ETFs at BlackRock, about the firm's 2026 thematic outlook, including:

• Why the AI infrastructure boom is still in its infancy.
• How thematic ETFs can give retail investors more precise exposure than traditional sector funds.
• What the rise of agentic AI, physical robotics, and tokenization means for your portfolio.

To catch full episodes of all The Motley Fool's free podcasts, check out our podcast center. When you're ready to invest, check out this top 10 list of stocks to buy.

A full transcript is below.

This podcast was recorded on May 31, 2026.

Jay Jacobs: Token consumption last year grew 17 times. Not 17%, which I think most people would view as a pretty good growth company, 17 times growth of token consumption. Essentially, as much money as the major large language bottle providers are plowing into capital expenditures, they can't keep up with AI demand. Even just in the last several months, I think the narrative has shifted in the market from that of we worried companies are over-investing in capex to what if companies are actually under-investing in capex?

Rachel Warren: That was BlackRock's U.S. head of equity ETFs, Jay Jacobs, breaking down what the data actually says about AI's growth trajectory. I'm Motley Fool analyst Rachel Warren. I sat down with Jay to dig into BlackRock's newly released 2026 thematic outlook, covering everything from the AI infrastructure build-out to tokenization to what retail investors should be doing with their portfolios right now. Enjoy. Hello, everyone, and welcome back to Motley Fool Conversations. I'm Motley Fool analyst Rachel Warren. Today, I'm excited to welcome Jay Jacobs, the U.S. Head of Equity ETFs at BlackRock, to the show. Jay oversees the overall product strategy, thought leadership, and client engagement for the firm's index and active equity ETF business. Prior to his current role, Jay founded and led Global x ETFs research and strategy team and previously served as a business analyst at the New York Stock Exchange, where he helped launch hundreds of ETFs on the NYSE ARCA trading platform. Today, we're going to be diving deep into the massive structural shifts shaping the global economy with BlackRock's newly released 2026 Thematic Outlook, which details how the next leg of AI Compute is colliding with physical power grid bottlenecks, surging sovereign defense spending, and a massive wave of real-world asset tokenization. Jay, welcome to the show.

Jay Jacobs: Thanks for having me on.

Rachel Warren: As U.S. head of equity ETFs, from your standpoint, I would love to hear your thoughts on how the view of a traditional portfolio has changed now that thematic funds have grown over 11X just in the past decade.

Jay Jacobs: I think it's important to recognize portfolio management techniques have always been evolving as the world has evolved as data and software has evolved to make portfolios be able to be managed in different ways and assess risks and opportunities in different ways. You go back to some of the factor research in the 1970s, the introduction of the style box in the early ‘90s, the GICS sector classifications that divvied up the world into different sectors in the late 90s. There's been a constant evolution of portfolio management. What we're seeing is one of the latest evolutions is really increasingly investors are looking at the world through a thematic lens. They see the rise of artificial intelligence, the changing demographics, the changing energy needs, the future of finance, as well as geopolitical fragmentation, all being major forces that are reshaping how they can think about risks and opportunities in their portfolio. As they assess those risks, they increasingly see how valuable thematic ETFs can be for fine-tuning their exposure to these themes in their portfolios.

Rachel Warren: Well, one of the things I wanted to talk about your internal model portfolios hit a 7.5% allocation, but the average moderate U.S. advisor model sits at just 3.6% thematic exposure. Your data actually shows that about 12% of analyzed U.S. advisor portfolios currently hold any thematic ETFs at all. I wonder if you could talk through maybe what's causing this gap, and does this mean that, sometimes we're seeing an underallocation to structural growth?

Jay Jacobs: Say there is an underallocation, or the way that people are getting exposure to these growth opportunities is through not always the most precise tools. I do think a lot of people out there think they're getting exposure to AI by allocating to the technology sector. In some ways, you are. Yes, the technology sector has exposure to names that are building large language models or building some of the important hardware that goes into data centers. But as we've also seen this year, the tech sector also has exposure to software names that have been disproportionately hurt by the rise of artificial intelligence and the risk that that presents to SaaS business models. I think what many people are learning in real time is just there's a difference between sector investing and thematic investing, and for some of these really disruptive themes, it takes a dedicated thematic ETF to be able to target them appropriately. We are seeing a gradual shift of more adoption of thematic ETFs amongst advisors. Yes, the average allocation is 3.6% as of our last reading, but you go back a few years ago, it was less than 3%. We're seeing a tick upwards. It's just somewhat lagging what we've seen in our own models which have more rapidly deployed thematic exposures given this market environment. I would expect this growth in advisors use of thematic to continue, though in the coming years.

Rachel Warren: Well, switching gears completely, we have to spend some time talking about AI, which obviously was something that was a really significant focus in the report, which I found incredibly interesting. What do I want to start with, market skeptics scream that tech companies are overspending on AI. We keep seeing those capex figures multiply. But what was interesting was your report shows that in the U.S., GEN AI infrastructure spending is just about 0.8% of GDP, compared to say, 4.5% for U.K. railroads in the 1860s, about 2% for US electricity, if you go back to the 1920s. Should one take away from this that the physical AI build-out is actually in its infancy? What are these numbers telling us?

Jay Jacobs: That's exactly right. On the scale of other major transformational events within the United States, AI capex has still not reached the upper echelons of that type of investment. Part of it is we're early. This AI boom has really only started since the end of 2022. We're a few years into it. We're seeing some of these capital expenditure numbers really accelerate upwards at a tremendous rate. I think we're going to see that percentage of GDP invested in AI continue to rise over the next several years. But the fact that it's still below what we saw as investment in railroads, investments in automobiles from a historical context just shows we're early. This country has been through transformations before. It's taken a tremendous amount of investment in each of these transformations. But the impact of those transformations can span many decades, as we've of course seen with the automobile, as we've of course seen with telephones. It's a reminder that we're early and it's still going to play out over the next several years.

Rachel Warren: Well, it's interesting to think about, as well, because you go back to, say, the telecom boom, in the 90s, that spent about 1.5% of GDP before crashing. Obviously, GEN AI spending is sitting about half of that right now. It's not a one-to-one comparison, either, but I'm curious what structural protections say prevent AI infrastructure from suffering dangers of overcapacity crash as we have seen with past buildouts.

Jay Jacobs: Frankly, I think a lot of this build-out is just a lot less speculative because so much of this compute that is being built out is almost instantaneously being monetized because of AI demand. What we show in the report is that token consumption last year grew 17 times. No 17%, which I think most people would view as a pretty good growth company, 17 times growth of token consumption. Essentially as much money as the major large language bottle providers are plowing into capital expenditures, they can't keep up with AI demand. Even just in the last several months, I think the narrative has shifted in the market from that of we worried companies are over-investing in capex to what if companies are actually under-investing in capex? Could we start to see bottlenecks and artificial intelligence, where some of the most powerful models frankly, have to be throttled because there's so much demand to use them versus the compute that's actually available across the economy. Yes, the capex is accelerating. The numbers are quite staggering of what we see being invested each year. However, the demand is backing it up, and the revenue from demand is immediately backing it up. This is not the same as speculatively building telecom infrastructure, and then, if we build it, they will come scenario. This is meeting real demand in real time.

Rachel Warren: Yeah, I think the other thing as well that I would like to dig into a bit more is this growth coming from agentic workloads, which, is essentially AI that can complete multistep tasks on its own. The report notes, this can increase relative to intensity by, a thousand times. We're seeing everyone from corporate America, the big tech companies, and beyond deploying AI agents. What parts of the tech stack can capture this exponential surge in data processing? Where are the beneficiaries? What can retail investors take away from that?

Jay Jacobs: It looks across the entire artificial intelligence tech stack. I mean, it starts with some of the lowest levels, which is really in the infrastructure to think about the power that's applying data centers, the data centers themselves, the real estate, the hardware going into those data centers. Think about all the semiconductors, whether it's memory, whether it's GPUs, whether it's CPUs that are powering those data centers. On top of that, there's the data layer. Think about the proprietary data that's training a lot of large language models. There's the large language models themselves that are being more and more powerful. We're seeing that software improve significantly year over year. Then of course, you have the applications and products that are using those large language models to utilize agents, whether that's, imagine having a financial analyst that can help you pour through news or earnings reports, sell-side reports, etc, consolidate all that information, put it into an Excel file or a PowerPoint presentation, you name it. There's a lot of things that AI agent can now be programmed to do and really take on a significant amount of tasks or people in a wide variety of different industries. That's why we're so focused across the entire AI value chain because as you see more adoption of agents, it's really going to flow across that entire value chain where you see companies profiting off of that.

Rachel Warren: There was data in the report from McKinzie that projected cumulative global infrastructure investment is set to top about $100 trillion by 2040, and that's driven by a range of factors, including, AI compute, national security, supply chain resilience initiatives. How can a long-term investor evaluate these sectors across this really, truly massive capital rollout we're seeing?

Jay Jacobs: Well, interestingly, despite the amount of capital we're seeing allocated to infrastructure, it remains a relatively small part of people's portfolios. In fact, average infrastructure allocation in the SP 500 is about only 3%, so less than some of the MAG_7 names alone. Yet we just see tremendous amounts of drivers for more infrastructure spending. We have changing demographics around the world, which is, growing economies, growing populations that need more infrastructure. We have aging infrastructure, particularly in the developed market, where a lot of it was built in the 1960s and needs to be refreshed. Have changing infrastructure demands, where it's not only about physical infrastructure, there's also needs for digital infrastructure going forward. There's really a lot of tremendous tailwinds behind infrastructure, and yet it remains a relatively small part of people's portfolios. I think we're going to see a significant amount of investment over the next several decades. I think a lot of that is going to increasingly come from the private sector, given that a lot of governments just simply can't afford to keep building more infrastructure, and that should likely drive more and more investors to allocate the infrastructure as an asset class in their portfolios.

Rachel Warren: I want to switch a bit to talk about the relationship between what we've been speaking of and tokenization digital assets. The report noted that the iShares Bitcoin Trust ETF became the fastest-growing ETP in history. It surpassed $70 billion in AUM in just 341 trading days across 2024 and 2025. What does that level of speed and adoption tell us about the current capital demand for digital assets?

Jay Jacobs: Well, IBIT was a product is a product that really bridges between traditional finance and decentralized finance. The idea that we could take a decentralized finance asset like Bitcoin, wrap it in exchange traded product and make it available to basically anyone with a brokerage account brought DeFi into the TridFi world, and we expect that trend to likely to continue. There's a lot of demand for assets that can behave differently than stocks and bonds. We've seen a tremendous amount of interest. From the traditional finance base in an asset like Bitcoin, where it's more driven by things like geopolitical uncertainty, rising distrust in institutions, the risk of debasement of currencies or rampant inflation, all of those things tend to be writing tailwinds for an asset like Bitcoin. We live in an environment where I think those are very real risks. Increasingly, very traditional portfolio managers are looking at Bitcoin as a way to hedge out some of those risks in their portfolio.

Rachel Warren: A significant share of tokenized, real-world assets currently reside on [inaudible] blockchain, and we're also seeing expectations that tokenization will continue to expand across asset classes. How do you see tokenization reshaping access, liquidity, and transparency for a broader range of investors?

Jay Jacobs: Well, it's likely to evolve. Right now, we largely see a tokenized cash or stable coins, and that's where the massive amount of volume is occurring today. Needs to be a market that develops around this. When you have tokenized assets, you need to have the infrastructure behind it. You have to have the market-making capabilities; there needs to be sensible regulation around it. There's a whole ecosystem that has to develop around it, but there's certainly the promise of tokenization that could allow for the 24/7 trading of assets, trading around the world, instantaneous settlement, perhaps easier access to decentralized finance tools like lending through SMART contracts. There's a lot of promise through tokenization, but it's also about really having an ecosystem develop around it to support it appropriately.

Rachel Warren: Couple more questions for you as we draw to the close of our discussion today. One, the 2026 outlook really did a brilliant job of connecting the dots between compute power grids and geopolitics and how all of these themes interplay. But looking beyond that, looking ahead to the next three to five years, what are maybe one or two emerging or under-the-radar themes or maybe tech breakthroughs that you think maybe investors should be paying close attention to?

Jay Jacobs: First of all, I would say, I think there's a lot of durability to the themes we talked about today. Yes, we call it the 2026 outlook, but in reality, these are things that we see multiyear, if not decades-long horizons behind. We are not trying to immediately pivot away from our interest in things like artificial intelligence or geopolitics or tokenization and beyond. What I will say is, I think the intersection of those themes and how they evolve in the next few years will be really interesting. One of the areas we did not talk about is the intersection of artificial intelligence and healthcare. This is one of the sectors that you could see both revenue acceleration through artificial intelligence. Think about developing revolutionary new drugs that hopefully treat various different diseases or ailments. But also, you could see cost-cutting benefits through artificial intelligence. Could it be faster with less trial and error developing those drugs that reduce the amount of cost to bring them to market? There's both a revenue and a cost opportunity in the healthcare space. Then we talked a little bit about it in the AI section as well. I think from just digital AI to physical AI with robotics with autonomous vehicles, that's something that we think is going to become increasingly important part of the conversation with AI going forward.

Rachel Warren: Well, and finally, what do you think are one or two important frameworks we should use to really filter out some of the short-term market noise and write out these generational mega forces over the long run?

Jay Jacobs: I think the important thing to look at is, what is the state of the technology? What's the use case? What's the size of the opportunity behind that use case, and then ultimately, what's the probability that it gets fulfilled? The earlier you are in a theme, potentially, the more opportunity you have, but also the more risk you have that it doesn't play out. Where we are with artificial intelligence today is really in a sweet spot where it's still very early. It still hasn’t seen economy-wide adoption and disruption yet. But we have enough evidence to believe that this is here to stay, that this is a real technology with many different use cases that continues to improve at light speed. When you combine those factors together, that's the conditions for a really important theme and potentially an important allocation in people's portfolios.

Rachel Warren: Fantastic. Well, I think you've given our listeners and viewers a lot to think about as we move ahead into the next decade of investment. Jay, thanks so much for joining me today.

Jay Jacobs: Thanks for having me.

Rachel Warren: As always, people on the program may have interests in the stocks they talk about, and The Motley Fool may have formal recommendations for or against. Don't buy or sell stocks based solely on what you hear. All personal finance content follows Motley Fool editorial standards and is not approved by advertisers. Advertisements are sponsored content and provided for informational purposes only. To see our full advertising disclosure, please check out our show notes. For the Motley Fool Hidden Gems Investing team, I'm Rachel Warren. Thanks for listening. We'll see you next time.

Rachel Warren has no position in any of the stocks mentioned. The Motley Fool has positions in and recommends Bitcoin, BlackRock, and iShares Bitcoin Trust. The Motley Fool has a disclosure policy.