Related Articles
Stay Up To Date
Something has gone wrong, check that all fields have been filled in correctly. If you have adblock, disable it.
The form was sent successfully
Artificial intelligence is becoming an energy and infrastructure story. As data center demand surges, power is emerging as a critical constraint and a potential source of long-term opportunity across generation, transmission, and distribution.
Transcript
Kevin Murphy: Hi. I’m Kevin Murphy. I’m a managing director at Sands Capital. Today, I’m joined by Brian Keegan and Jacob Kann. Brian and Jacob are both senior research analysts here at the firm. They cover a number of businesses across multiple industries, including industrials, space, defense, and power.
At Sands Capital, we remain committed to growth investing, but we also look for long-duration opportunities where durable businesses can grow through changing market environments.
Energy has clearly moved higher on that list as artificial intelligence (AI) drives a
sharp increase in electrical demand and creates potential investment opportunities
across the entire value chain.
Brian, Jacob, thanks for joining me today.
AI Driving Massive Rise in Power Demand
Kevin Murphy: Brian, let me start with you. Why has power become such a critical topic now?
Brian Keegan: Well, it’s a critical topic now because of the changing pace. If you
look over the last 15 years, U.S. power demand was almost flat and utilities were
mostly focused on renewable integration versus capacity expansion. That’s very
different than what we’re seeing now. AI is changing the power demand equation,
as the data centers required to support it require enormous amounts of continuous and dispatchable power. As a result, we’re moving from a multidecade period of flat load growth to an environment of mid-single-digit load growth, at least for the next few years.
I know mid-single-digit doesn’t sound huge, but the U.S. grid is a very large system, and the incremental power demand that we’re talking about each year is on par with some Western European nations. That’s really the issue here. It’s not just that power demand is rising. It’s that it’s rising very quickly and at a scale that will require enormous amounts of power infrastructure to address.
AI as an Infrastructure Story
Kevin Murphy: This is really an infrastructure story as much as it is a technology story?
Brian Keegan: Exactly. We’re seeing the digital economy overlap with real-world infrastructure more now than ever. The scale here is huge. Just to put this in context, we think over the next five years, AI could require up to 90 gigawatts of incremental power production capacity. One gigawatt is enough to power all the households in Washington, D.C. Think 90 Washington, D.C.s worth of power. It’s an enormous number.
Kevin Murphy: Jacob, when you look at that backdrop, how are you framing the opportunity?
Jacob Kann: We start with the bottlenecks. Power is accelerating. The key question is where the system starts to break down. Is the constraint generation, transmission, labor, or is it simply the time it takes to build new infrastructure? Once you identify those pressure points, you can start to see where the most compelling opportunities could be, and that’s how we’ve been approaching the research. We’re trying to understand which businesses are mission-critical to the buildout.
Kevin Murphy: When you look for those beneficiaries, where are you finding them?
Jacob Kann: One area is transmission and grid infrastructure. If electricity demand rises sharply, you don’t just need more power generation. You also need the ability to move that power where it’s needed across the grid. That’s a major challenge in the U.S. because the grid is aging, it’s fragmented, and in many places not designed for the level of incremental load demand that Brian described earlier.
To make an analogy to roads, when traffic is congested around cities, municipalities
start to build highways. The need to build large-scale, high-voltage transmission to
reduce grid congestion has a similar dynamic. If utilities don’t make these investments, they run the risk of blackouts or grid failure. That leads us to companies involved in transmission infrastructure. Those businesses may not always be the most obvious AI beneficiaries, but they can be some of the most important enablers of the broader ecosystem. A company like Quanta Services is a good example. It’s really the people behind the buildout. They have an employee base of about 70,000 skilled craft workers that play a critical role in building and maintaining power infrastructure across the U.S. In the backdrop we describe, Quanta’s capabilities should become increasingly valuable.
Grid Access and On-Site Power
Kevin Murphy: Brian, what are your thoughts on this?
Brian Keegan: Another area where we see a constraint is in grid access. Even if the utility has power available, there’s a lot of regulatory work that needs to be done to integrate these projects, given their size. Some data centers have addressed this by building power on-site. This is a model they call “bring your own power.” We think a major beneficiary of this trend will be Siemens Energy. The company’s gas turbines are increasingly deployed by data centers on-site, given their efficiency and high-power output. Data centers have gone from a negligible piece of that business several years ago to over 20 percent of the turbine order book today. We expect that number to continue to move higher as AI continues to scale.
Kevin Murphy: Siemens is a fairly well-known business, and gas-powered turbines are also well-known technology. Jacob, where are you finding opportunities in maybe less well understood spaces?
Jacob Kann: Another company we’ve spent time on is Bloom Energy, which also fits the “bring your own power” trend. Bloom’s fuel cell technology has become especially relevant in a market where speed to power is critical. Think of fuel cell technology as a dry battery, constantly refueled by natural gas. The systems are packaged in modular units called Bloom Servers. They’re the size of a large refrigerator, and importantly, they’re being used as baseload, always-on power for data centers, not just backup. If grid access is delayed and traditional turbine availability is constrained, then alternative and faster on-site solutions can become more important. Bloom is a different type of investment case than Quanta or Siemens. It’s more technology-driven, and it comes with a wider range of outcomes. But if customers increasingly value modularity, deployment speed, and reliable on-site power, we think Bloom could become an important beneficiary of that shift.
Power as a Secular Growth Opportunity
Kevin Murphy: You’ve talked about some innovative companies. But at the end of the day, we’re really talking about energy power often associated with commodity or utility businesses. One question that might come up from clients is whether embracing names like these represents a style drift away from growth for Sands Capital. How would you respond to that?
Jacob Kann: We don’t see it that way. What’s changed is that there’s an external catalyst, in this case AI, that’s really transformed these companies’ opportunity sets. In our view, that makes them more secular and less cyclical in nature. This is not about drifting from style. It’s more about recognizing when the business we understand well is benefiting from a meaningful change that can make its growth trajectory more durable.
Kevin Murphy: Brian, how do you see this? Is this a short-lived trend, or is there something more secular and durable?
Brian Keegan: We think that it’s likely to be much more durable and much more long term. Infrastructure takes a long time to build. It took us over a century to build out our current grid, and we’re now moving to an environment where we’re expanding it rapidly and off a very large base. The future of AI is also likely to extend the duration of these power trends. In addition to broader growth in inference or larger models, things like moving from text generation to more advanced reasoning, image and video generation using AI models, and agentic demand should continue to drive power consumption into the future. These things can be significantly more power intensive on a per inference basis than what we’re using today. So while I think there’s always going to be ebbs and flows in different parts of the value chain, it’s a big industry. The broader backdrop should be one of a multiyear investment process across the electrical system, with a long duration thereafter.
There’s More to AI Than Semiconductors
Kevin Murphy: Before we wrap up, I would love to get your final takeaways. Brian, let’s start with you.
Brian Keegan: My biggest takeaway here is that AI is not just a semiconductor and software story. It’s also an energy and infrastructure story. If AI continues to scale, productivity at a country level may be increasingly tied to the amount of power that we can generate on the ground. It’s a pretty exciting time for the industry.
Kevin Murphy: Jacob, any unusual insights from your view?
Jacob Kann: One conclusion is we’re going to stay curious and open-minded as facts on the ground change. This is a fast-moving market, and our job is to keep learning as the opportunity set evolves. As we laid out today, we’re excited about what we’re seeing.
Kevin Murphy: That’s a great place to end. Energy may not be the first thing people think of when they think about AI, but it may become one of the most important. Brian, Jacob, thanks for joining the conversation.
Disclosures:
The views expressed are the opinion of Sands Capital and are not intended as a forecast, a guarantee of future results, investment recommendations or an offer to buy or sell any securities. The views expressed were current as of the date indicated and are subject to change.
This material may contain forward-looking statements, which are subject to uncertainty and contingencies outside of Sands Capital’s control. Readers should not place undue reliance upon these forward-looking statements. There is no guarantee that Sands Capital will meet its stated goals. Past performance is not indicative of future results.
All investments are subject to market risk, including the possible loss of principal. Recent tariff announcements may add to this risk, creating additional economic uncertainty and potentially affecting the value of certain investments. Tariffs can impact various sectors differently, leading to changes in market dynamics and investment performance.
As of April 9, 2026, Bloom Energy, Quanta Services, and Siemens Energy were held across Sands Capital strategies. The specific securities identified and described do not represent all of the securities purchased, sold, or recommended for advisory clients. There is no assurance that any securities discussed will remain in the portfolio or that securities sold have not been repurchased. You should not assume that any investment is or will be profitable. The companies were chosen based on an objective criteria and represent the power and energy companies held in Sands Capital public strategies.