Key Sections
Fueled by AI and cloud computing, the rapidly increasing demand for digital infrastructure is putting more shovels in the ground than ever before. In the first of a series of articles featuring Blue Owl's analysis of key topics in the global economy, Matt A’Hearn and Gary Rozier of the Real Assets platform discuss the opportunities for data centers.
AI is expected to add up to $15 trillion to the global economy in the next five years.
The need for more data centers could result in $1 trillion of new capital expenditures.
Big tech firms are spending $300 billion on data centers this year alone.
How would you describe the data-infrastructure landscape right now?
Matt: I believe we are seeing a generational opportunity to invest in this sector—just given the scale of it, and the tailwinds associated around digital infrastructure and data centers in particular.
We talk about what’s happening right now as a “wave on wave” of industry drivers: It really started for us about seven years ago, when cloud compute began rapidly growing—companies like AWS, Microsoft, and others whose cloud businesses grew tenfold since that time. In November of 2022, when ChaptGPT was commercially launched to the world, the second wave came, which was AI. And I think what we're seeing out of AI is the enormous potential it brings. For example, there are estimates that it could add $15 trillion to the economy over the next few years, and we view the infrastructure that's necessary to support it as a trillion-dollar market opportunity.
There’s more than one kind of data center, and there’s more than one kind of data center user. Where are you focusing your attention across the digital infrastructure landscape?
Gary: Specific to our net-lease strategies, our focus is primarily targeted on transacting with larger hyperscale technology companies as an equity partner, securing financing to fund development of a data center that is pre-leased on a long-term basis, and ultimately owning the stabilized asset. As such, we tend to be later stage and more narrow in scope with respect to the type of data center end users that we are working with and the type of projects we engage in.
Matt: Hyperscale companies are the businesses that are primarily driving investment into AI. Just to give you a sense of what I’m talking about: The combination of Amazon, Microsoft, Google, Meta, and Oracle are spending over $300 billion of cap ex this year alone2. Along with these companies, newer AI businesses like OpenAI and Coreweave are increasingly important players in the industry.
Power availability is a big issue in data centers, given they use so much of it. How does that play a role in determining where you want to build?
Matt: Power has become, to some extent, the biggest bottleneck for the industry. Power usage in the US has been relatively flat—maybe it bounces up and down a little bit over the last 10 years, but power usage has been stable. Now we're seeing a meaningful upswing in power usage, and this is expected to continue.
This growth has been driven not only by AI (and ultimately data centers) but also increased manufacturing and the further electrification of our lives through products like EVs. We haven't matched increased demand with increased supply. Additional investment and production needs to be put in place with power generation and transmission infrastructure.
That imbalance is playing out differently in different markets: What might be creating challenges in Phoenix, for example, might be different than what's happening in Virginia. You really need to have a good ground game to understand the issues and work with all stakeholders to help come up with solutions. A few years ago, you could buy a nice rectangular piece of dirt in Virginia and say, “The site is perfect, and the utility will provide me the power I need because they're a utility and that's what they do.” Now, if you do that, you could be looking at four, five, six years before you actually have the power.
We saw this dynamic playing out early in the sector and made a conscious decision to build a team internally with expertise in this area. This investment in the team would provide us with the resources to work with our partners to solve these problems. You really have to be smart about this topic—there are pockets of power, but it's hard and you need to understand how to navigate this.
Gary: It also creates a tertiary opportunity as well. In addition to the need for capital for the data center itself, there are also capital needs for other areas of the infrastructure buildout—power being one of the largest. That's another area where we can help supply the necessary financing, either to the hyperscaler and/or developer who may be responsible for delivering power. Furthermore, we expect increased growth within semiconductor R&D, manufacturing, and accelerated onshoring. We view the scarcity of capital as one of our greatest areas of opportunity.
Infrastructure like what you’re describing is a long-term proposition, but technology can move rapidly, such as the recent news about DeepSeek’s potentially cheaper, more efficient AI model. How do you account for that kind of tech obsolescence risk in your strategy?
Matt: To some extent, we need to have improved efficiencies in the marketplace, or the model might not work well. And so we think it could be a positive for us (and the market) because if you get improved efficiency and lower costs, you're likely going to get a higher use case—which means higher demand—which we can benefit from.
Even if you cut my trillion-dollar number in half to $500 billion, this still reflects massive growth in the industry. Also, we are not really trying to pick winners and losers in AI—the technology is going to continue to change. We'd rather play at the infrastructure side and not have to be as sensitive to who wins at the end of the day.
Gary: We’ve seen this evolution in various areas of technology and other industries over the years, where the development of lower-cost or more affordable models has expanded production and accelerated the growth of what is brought to market. The outcome is often an expanded market and wider adoption. We see this to be true in AI as well and the need for more digital infrastructure spending to continue.
Matt: We've used this analogy for a very long period of time, but I think it still holds: If you think about a flat screen TV, it's going to change every couple of years. The technology is going to change; the picture will get better; the TV will get thinner. It's also going to decrease in cost, and that might mean you buy more of them. But even with these changes, it will always need power and connectivity. To some extent, that is what we are providing in a data center—the plug and the internet connection. Just like TVs, we are not getting into the technology side of the equation like servers. We're more focused on the longer-term infrastructure which all TVs (and servers) will continue to need.
How is financing evolving to meet this need?
Gary: We are primarily providing capital for the building itself, not necessarily for the IT that goes inside. It is often the preference of the tenant to structure their spending on what is occupying the asset, as that’s where their revenue is ultimately derived, which we believe is a more accretive use of their capital. That can be true in data centers, as it can be in a manufacturing facility, or in a chain of essential retail stores. We assume the cost of the shell so our hyperscale partners can then focus on the revenue side of the asset.
Interviewees
Matt A'Hearn
Matt A'Hearn is a Senior Managing Director at Blue Owl and the Head of Digital Infrastructure. In his role, he is responsible for leading the overall management of the Digital Infrastructure platform, including strategy, investments and portfolio management as well as other corporate initiatives. Matt also serves as Chairman of the Digital Infrastructure Investment Committee.
Gary Rozier
Gary Rozier is a Senior Managing Director at Blue Owl and member of the Real Assets Management Team. In his role, Gary serves on the investment committee and is responsible for reviewing and approving Real Assets transactions, as well as interfacing with limited partners on the investment portfolios.
Sources
Important Information
Past performance is not indicative of future results.
This article is for informational purposes only and is not an offer or a solicitation to sell or subscribe for any fund and does not constitute investment, legal, regulatory, business, tax, financial, accounting, or other advice or a recommendation regarding any securities of Blue Owl, of any fund or vehicle managed by Blue Owl, or of any other issuer of securities.
The views expressed and, except as otherwise indicated, the information provided are as of the report date and are subject to change, update, revision, verification, and amendment, materially or otherwise, without notice, as market or other conditions change. Since these conditions can change frequently, there can be no assurance that the trends described herein will continue or that any forecasts are accurate. In addition, certain of the statements contained in this material may be statements of future expectations and other forward-looking statements that are based on the current views and assumptions of Blue Owl and involve known and unknown risks and uncertainties (including those discussed below) that could cause actual results, performance, or events to differ materially from those expressed or implied in such statements. These statements may be forward-looking by reason of context or identified by words such as “may, will, should, expects, plans, intends, anticipates, believes, estimates, predicts, potential or continue” and other similar expressions. Neither Blue Owl, its affiliates, nor any of Blue Owl’s or its affiliates' respective advisers, members, directors, officers, partners, agents, representatives or employees or any other person (collectively the “Blue Owl Entities”) is under any obligation to update or keep current the information contained in this document.
This material contains information from third party sources which Blue Owl has not verified. No representation or warranty, express or implied, is given by or on behalf of the Blue Owl Entities as to the accuracy, fairness, correctness or completeness of the information or opinions contained in this material and no liability whatsoever (in negligence or otherwise) is accepted by the Blue Owl Entities for any loss howsoever arising, directly or indirectly, from any use of this material or its contents, or otherwise arising in connection therewith.
Copyright© Blue Owl Capital Inc. 2025. All rights reserved. All individual charts, graphs and other elements contained within the information are also copyrighted works and may be owned by a party other than Blue Owl. By accepting the information, you agree to abide by all applicable copyright and other laws, as well as any additional copyright notices or restrictions contained in the information.