Beyond tech: harnessing AI’s potential for innovation and growth

 Jennifer Martin

Welcome to the Angle from T Rowe Price. The podcast, where curious investors can gain an information edge on today's evolving market themes. Sharper insights on the forces shaping financial markets, begin here. In this season of the angle, we are focusing on the exciting world of artificial intelligence, otherwise known as AI. I'm Jennifer Martin, a global equity portfolio specialist, here at T. Rowe Price Associates.

And I will be your host, as we learn more about this intriguing, and rapidly evolving area of the global economy and financial markets. We'll investigate the technology's rapid ascent, impact on different economic sectors, its potential within investment research and the emerging sustainability challenges surrounding its adoption.

In today's episode, we take a deeper look at how AI is being utilized, and some of the early impacts we are seeing at a sector-specific level, specifically within the basic resources and real estate areas.

Joining me for this discussion are two experts in these respective fields. Rick de los Reyes, basic resources analyst at T. Rowe Price. And Gregg Korondi, T. Rowe Price’s real estate sector analyst. Welcome to the podcast Rick and Gregg, it's great to have you both on today's episode.

Gregg Korondi

Thanks for inviting us, Jennifer.

Rick de los Reyes

Yeah thanks, Jennifer. Great to be here.

Jennifer Martin

I'm interested to know when each of you first heard of AI being applied in your respective sectors.

Gregg Korondi

Sure. Yeah. For me, it was 2009 coming out of the GFC. People were still talking about cloud adoption and what came after that. And the context of the idea from the data center providers was that the next stool, the next leg of the stool would be the Internet of Things and machine learning and AI. And so, thinking about how people integrate manufacturing processes, how they integrate driverless cars, you know, the consumer side was toasters and refrigerators and aggregating the data and processing things quickly. And AI and ML – machine learning – were kind of being used interchangeably at the time.

Rick de los Reyes

Yeah, I follow a lot of commodity companies and oil and gas and mining companies that are in extractive industries. And I'd say they've been thinking about this for, you know, probably close to a decade. Certainly, they've been trying to move towards automation for a very long time. And I think a lot of that was a predecessor to using AI technology.

I think probably the earliest adopters were, if I think about the iron ore miners in the Pilbara region in Western Australia, where they for a long time now have been using driverless trucks and autonomous vehicles. And, you know, I think if we're really being honest and we talk to those companies, they would tell you that a lot of that was really out of just the necessities of the labor situation.

And it gets very difficult to find, to find labor and to find people to do that job. You think about where those mines are – they're kind of in the middle of nowhere. They’re out in Western Australia. There's no cities or towns nearby. To get to work there you basically have to live in some other part of Australia and then they fly you in for a couple of weeks at a time and you live on site for a couple of weeks, away from your family, you know, living and doing a dangerous job. And then when your two-week shift is over, they fly you back home for your two weeks off.

And the reality is that people don't want to live that way. And so, it's getting very, very difficult to attract people to those jobs. And the amount that they were having to pay people was getting, was getting pretty prohibitive. And so, to be able to just use autonomous trucks and not need as many people on site was certainly a good way for them to combat that. And when you think about it, companies that are in extractive industries, in commodity type industries, are kind of the natural customers for something like AI because they're companies that can't compete on price or on product, right?

They're producing something that's a commodity – it’s the same product that somebody else is producing. So, you know, if I'm extracting natural gas from the Permian, the guy next to me is producing the exact same natural gas, and we're getting the exact same price for our natural gas. So, the only way that I can compete with that guy is by doing it for a cheaper price, right?

By extracting more efficiently and for a lower price, so that if that natural gas price goes down, he's the one who has to shut down production, not me, because I'm the more efficient, lower cost, producer. And so, to the extent that you can use technologies like artificial intelligence to lower your cost of production, that's how you compete in the commodity world. And that's why you're seeing companies that provide services, they're investing a lot of time and money and energy into trying to develop these technologies so that they can provide more efficient solutions for their customers.

And many of them, they’ve put out press releases already, talking about some of the successes that they've had with early testing, with AI, and, in some cases, they're increasing drilling speeds by 40 to 60%. And that's huge for someone who's trying to be a

more efficient producer and lower their costs in an industry where it's all about competing on costs.

Jennifer Martin

Those are very useful insights because I think for the basic resources sector, this is maybe not the first sector that springs to mind when we're speaking about artificial intelligence. So, one of the things that you did bring up was related to how AI is really working on the cost optimization. So, let's take this cost optimization idea a little further. Any time a machine fails, or if there is any downtime in production, this must end up adding to overall costs. Can AI help improve efficiency in this regard?

Rick de los Reyes

Yeah, absolutely. And one of the, you know, one of the AI buzzwords that you hear talked about in the industry is is is predictive maintenance. And so, the idea of predictive maintenance is that you use AI throughout your sort of supply chain to be able to predict where production issues might come up, because absolutely, if you’re a commodity producer, you know, downtime is your enemy, right?

Time is money. And any time that you have to take downtime because of equipment failure, you know, that comes right off of your top line. And so, you know you think about the example that I gave before of somebody pumping gas out of the Permian Basin. You know, think about that supply chain and how long it is from the time I'm getting gas out of the Permian.

I've got to get that to a pipeline, then from the pipeline it has got to go into the storage facility. From a storage facility, it goes through another pipeline until it gets to a gas fired power plant. That gas fired power plant turns it into electricity, which then has to go into a distribution system, which then goes to the transmission system that goes all the way to your home and eventually ends up in the power outlet where you're plugging in your lamp.

So, think about how long that that supply chain is and think about the number of things that can go wrong along that chain. So, to the extent that you can have sensor technology and AI employed all along that chain to let you know where the potential weaknesses are, to let you know where there's potentially going to be a maintenance issue, before it happens, then that does wonders to increasing your efficiency and, once again, increasing your top and your bottom line.

Jennifer Martin

Rick, you mentioned previously that working in the natural resources industry involves a lot of hard and dangerous work in fairly, in fairly far-flung locations. I know that safety issues are also a primary concern for basic resources companies. Can you elaborate on how AI is being utilized to improve worker health and safety?

Rick de los Reyes

Yeah, look I think that the potential for AI to make a really big impact on safety is is is, you know, really one of the huge advantages and really, really, one of the big use cases for it. And there's no question that what we look at are some of the most dangerous jobs in the world. And when I started out in the mining industry analyzing that sector many years ago, I traveled to South Africa. I spent time down in the platinum mines. And platinum mining in South Africa is one of the most dangerous jobs in the world. It it it involves mining extremely narrow veins. So, you've got to put human beings into very small spaces underground for them to mine very narrow veins and it's tragic – but far too frequently there are rockfalls that that involve fatalities, and that's certainly nothing that anybody wants to see happen. And so, to the extent that we can automate that process and mechanize that process and use AI so that we can remove human beings from even being in that dangerous position in the first place, that’s certainly something that all of us want to do. And I'd say it's not just the humane aspect of it, right?

There's, of course, a humane aspect that none of us want people to get hurt and certainly don't want to see anybody getting killed. But there is also just a cost aspect to it as well. So, there's economic reasons to do it. Because any time you do have an accident, you have to shut down that operation. And, once again, downtime is just lost time and time is money.

And so, nobody wants to see that downtime, that downtime happen, right? So, companies are definitely highly incentivized to keep human beings out of harm's way so that they never have to take those kind of safety downtimes. You know, if I think about the example that I gave before with autonomous trucks that were used out in the Pilbara in Western Australia, you know, the advantages are not just that it frees up your labor, but it's also a safety advantage, right?

Autonomous trucks don't have to pass a drug test, right? They don't need eight hours of sleep every night. They don't try to work through a sickness or a cold or whatever else. All those things are safety issues, and safety issues not just for that driver, but for the people who are working around him or her. And so, to the extent that you can avoid having to even worry about a human being falling asleep at the wheel, that's going to improve safety and make the entire workplace safer for everyone.

Jennifer Martin

Rick, thanks for your insights. I think we can appreciate the complexities of the natural resources sector and hopefully the improvements that artificial intelligence can have on the industry. So, switching to the real estate sector, again, not the most obvious area of AI application, but certainly an area of substantial data generation. Gregg, can you provide some insights on how AI is being applied within the real estate field and some of the early impacts you are seeing?

Gregg Korondi

Sure. I think just starting with the most obvious candidate for where AI is impacting the real estate world is the home of the AI compute functions, and that being the data centers. So,

for a long time we've seen clustering of data center compute and connectivity. So, I've always thought of the data center sector as two different functions. One is a large compute function, and one is a high connectivity function. And the need for them to be in the same spot is helpful, if possible, but not exactly necessary. So, latency requirements had people choosing spots with high density of data center demand and high density of power and connectivity, in one location. The large compute function has always had an ability to be a little bit more ubiquitous and where those locations end up being.

The understanding the difference about what needs to be where, and when, is a really interesting dynamic to see how it impacts real estate. I’ve always had the view that the larger compute functions, are a little bit more - there's a lot more substitute-ability. Twenty years ago, there weren't a lot of people that knew how to build third-party data centers to host different tenants within a building. Most people built their own data centers for themselves, and they were very purposely built and sometimes not highly reusable.

So, what does AI mean for that property type? You know, I think the natural inclination is people think it's going to migrate to where this availability of power and so will be easier to get because the infill clustering markets have run out of either land, at a reasonable price. They've run out of some of the connectivity for the power, or they've actually just ran out of power.

So, as the power and demand needs for some of the functions increase, where they end up locating will be an interesting cross-section of how this demand translates into real estate. What we've seen so far is people have just accelerated what they think they'll need over the next five years in the markets that they are most focused on.

But we're now starting to see early signs of people locating next to nuclear reactor plants, in central Pennsylvania, they're looking at Wisconsin, they're looking at Michigan, North Dakota, markets that are not near data usage, right? So, the latency issue is becoming less important and now the power generation is. The final part of that is trying to understand how vast this demand is going to be. The common assumption is because GPUs can do significantly more computational functions at six to eight times the amount of power density, that's just the straight math. What we've seen in some of the inferencing applications is that the utilization take up is much slower. So, the locations close to the rooftops where communication is paramount over power, are less intensive, so far. It's still early, so, we'll see.

The large language models, the heavily compute intensive stuff are being located further out and they're starting to migrate. But there hasn't been this seismic shift of everything's moving out. Atlanta signing a gigawatt lease deal, for a market that has 600 megawatts of power, is massive. It’s not moving out to the middle of a very rural area that might have cheap power and might have easy access to build. So, understanding how that manifests in the data center world I think is going to be extremely interesting to to watch and try to figure

out who has pricing power. Part of the dynamics around real estate is that the location kind of drives the power for asking rents, and how that algorithm flows through the cash flows.

Right now, those cluster markets have a lot of pricing power and it's a very interesting time, and lots of people willing to pay a price.

Rick de los Reyes

I think that's an important point because it's not just about the power, but it's about transporting that power. And the minute that you try to transport any power across state lines, you increase a lot, you increase the level of complexity quite a bit. And so that's why I think your point about places like Pennsylvania are probably correct. You want to go to places where you can get the power directly and have it delivered directly to your data center without having to go through a long transportation, you know, supply chain.

Gregg Korondi

Yeah. And the key around that has always been the latency discussion. The reason that it hasn't migrated out further is people want to be in the six-millisecond latency zone. And if the idea is that the programing is linear, and you don't need multiple connections, you don't need access to every telecom provider, every cloud provider, you can kind of just access at one point and run a fiber line out from an adjacency that's already going through a major hub. It's a lot cheaper than doing the power and trying to run things that way. It's actually a lot less detrimental to the timeline of activating something. So, there's this giant feeding frenzy of as much power as possible. As soon as humanly possible. So, thinking about the different markets, there's a queue forming for when people can get access to this power.

And if you can go to a market where the queue is only two years, you're now ahead of all these other markets. So, as long as the latency needs aren't being compromised, which are a lot easier to solve, the ‘where’ matters less. Now the flip side of that is will the ‘where’ matter less ten years from now? Right? Like the third-party data centers are very reticent to build a data center in the middle of nowhere to a single tenant, for a specific purpose. Because, once they get past year three of the lease, when there's less than five or six years remaining, the residual value is unknown. If that tenant doesn't renew, they might not recapture their investment. So, they have to think about, how liquid these options will be and what they're building in terms of how much power density, which is an interesting and burgeoning question right now.

Rick de los Reyes

But aren't a lot of the data centers being built by the hyperscalers themselves? So, Google, Microsoft, Amazon, so they are the customer, right? So, for them, it should be easy to locate it wherever is the most efficient place to do it, wherever they can find the best power source, right?

Gregg Korondi

They are the most natural builders of the AI data center, right. So, like if we think about what the third-party data centers want to build, they don't want to build a really small building

with an incredible amount of power density, in the middle of nowhere. That has to kind of go to the balance sheets of these hyperscalers that have the ability to pay for it.

It's been interesting because you’ve seen a trend. Initially, people were reluctant to use third-party data centers over time and they've migrated into that modality over time, but particularly enterprise users, because they realize that they can't build a better data center, they can't operate a data center. So that is a good utilization. And even the hyperscalers were leaning heavily ex-U.S. to build data centers or to lease data centers from people that were building them third-party. I don’t; I think that's got to shift back. I think, and you're seeing it, the hyperscalers are negotiating directly with local municipalities, local states. People are now land brokers for data centers. And you also want to think about what you're building, right? Like. those third-party data centers, so far, they're building the same data centers they were building five years ago.

So, nobody is really leaning into that except maybe the hyperscalers. I haven't seen a lot of evidence because most of what they're signing hasn't been built for this yet. So, if they had data center capacity, it was already built to the old specification. But the dynamic of building a GPU data center, I don't think is where this ends.

Jennifer Martin

Well, let's continue on the discussion of of, really, power utilization, I mean an area that's being massively impacted by the growing utilization of AI. This clearly has major implications for the basic resources sector. But the knock-on real estate impacts that you've also already really highlighted, Gregg, they are significant. Can you explain how the growing demand for power is likely to impact your respective sectors?

Rick de los Reyes

Yeah, look, I think the demand for power here in the U.S. is going to come from these data centers is probably something that's not getting enough attention from policymakers, it’s clearly getting attention from the hyperscalers that we talked about. I think people like Google and Amazon are already trying to secure power directly with utilities. But look, the total amount of power that's going to be needed, you know, obviously is just anyone's guess at this point.

There are a lot of estimates out there. I've seen estimates that the total amount of power is equivalent to, you know, what's used in entire European countries. Right? And so, one way or another, it's going to be more power than we're using right now. And that by itself is something that we're not used to because electricity demand in the U.S. has actually been largely flat for about the last 15 years or so. And so there hasn't been a lot of thought put into adding new generation capacity. And there's going to have to be thought put into that now. Right, because electricity demand is going to grow to some extent, right? And these things do use a lot of power. I mean, these things are the size of an Amazon warehouse, but they're three stories tall. Right. The amount of power that they use, 

Jennifer Martin

And they’re noisy.

Rick de los Reyes

And then the amount of power they use is measured in gigawatts, you know, not megawatts. So, they do use a lot of power. And we've got to think about where that power is going to come from on a grid that, by the way, we know already experiences times of stress now. Right? We've seen it already in periods of extreme weather in the U.S. where there's there’s risk of power failures already. And so, how do we do that? Where's that power generation going to come from? Well, we can look at it through a process of elimination. Right? So, it’s not going to come from coal, right? Because we’re phasing out coal and there doesn’t seem a lot of appetite to change the phasing out of coal that’s going on in this country right now, obviously, for environmental reasons. So that’s out. It can’t come entirely from renewables, renewables like wind and solar. They can be part of the solution, but they can’t be all of the solution because of intermittency, right? So, wind power only works when the wind is blowing and solar only works when the sun is shining.

And, you know, one issue with data centers is they have to run 24 hours a day, right? They don't get to take downtime. And so, you need to have reliable base power. And, you know, wind and solar can't really provide that by themselves. So, then we look at nuclear. Once again, I would say nuclear could be part of the solution, but it can't be all of the solution. And certainly, the lead time to building any new nuclear capacity would be extremely long, and we're going to need this electricity sooner than that's going to be able to provide.

So, when you kind of go through the process of elimination, you're left with natural gas. Right now, the United States is is fortunate to have very inexpensive sources of natural gas. And the reality is that if we're going to need reliable baseload power to power these data centers, it's going to have to come from natural gas. And so, this gets to the point that Gregg and I were talking about earlier, about where do you locate these data centers?

Well, if I were looking to locate a data center, I would want to locate it some place that's close to a cheap and reliable source of natural gas and preferably without crossing state lines. And so, you know, Pennsylvania's a good choice because the Marcellus is there. Texas with the Permian, the Haynesville - areas like that, where we have natural gas that we know that we can get to, those are the types of places that I'd be looking in order to secure the electricity demand that we're going to need for these.

Gregg Korondi

Rick, what's the view from the natural resources guys about the small modular reactors, the SMRs?

Jennifer Martin

And maybe you could explain what an SMR is.

Rick de los Reyes

Sure. Good point. SMR is a small modular reactor, which is essentially a small nuclear generating capacity. It’s a technology that is fairly new. There's a lot of promise to it, a lot of ideas about using it. But, once again, I think that it's a little bit of a ways off, right? It's not something that's immediately available for us to use. And so, I think, in the immediate term, in order to achieve the electricity generation that we need, it is going to have to come from gas. If we can get SMRs to really work in the future, then that has a lot of uses in the future, right? Not even just data centers. So, the promise of SMRs is definitely very high, but it's something that's further in the future, I think right now.

Gregg Korondi

Yeah, I mean, that's interesting because the way I think about how this opportunity set is unfolding is still locationally based, where are these going to be? What's it going to look like in ten years? And if SMRs are a real solution, that means that Montana facility that we're talking about, or North Dakota, that might not be worth much because you might be able to locate it three miles down northern Virginia and pop up, put it in West Virginia, just over the state line.

Jennifer Martin

Gregg, do you want to spend a little bit more time explaining how the growing demand for power is likely to impact your respective sector?

Gregg Korondi

Sure. Yeah. I mean, I touched on it a little bit. Well, there's two things that I’m really focused on. One is the pace at which they take down the available power supply and the prospective supply. Right now, it feels a little bit like a pie eating contest. Everyone's looking at the backlog of requests that are in the queue from these power companies and saying, Holy cow, I'm not going to get what I want. We're all going to get cut back. I'm going to put in for more sooner than I think I need it. And then figuring out how much to draw down because the monetization of these, the technology to justify the investment in these data centers is still being crystallized.

So, right now, people are worried they won't have the capacity they need in the timeframe they need it to win. Right, because there is another argument here too right. Like, does everyone have the equal right to win from AI? Meaning do we need 17 search engines powered by AI algorithms? Right. Like, I'm not sure. I don't know the answer, but probably not. So, if people are all building dual paths and saying, I need this capacity for this application, but there's only going to be one or two winners, there could be a scenario where it just takes longer to absorb what's been requested. So, we can, as we think more about kind of what ancillary knock-ons there are in the real estate sector, I think there's some interesting topics to cover there too.

Jennifer Martin

I know real estate is an incredibly data intensive sector, and so I know AI is probably a phenomenal tool to have in analyzing some of the analytics that you just highlighted. I don't know if you want to spend a moment on that.

Gregg Korondi

Sure. Yeah. I think it'll help crystallize speed to market by understanding the local dynamics of that market. As much as we talk about real estate in generalities, the reality is it's a very local specific business, it comes down to knowing a market very well. And there's always been something they call the dumb tax, right? Like if you're owning a group of assets in New York and you buy in Atlanta, you generally aren't as familiar with the political environment, the demand drivers, and the kind of geographic specifics of a different sub-market.

So being able to pull in all the data from the data providers quickly, assess the differences of how things are correlated and how they are not, will help in selection, it will help understand who the tenants are to the extent that there's overlap, they can leverage existing relationships. So, I think those are all kind of incremental changes, and we have some building blocks of those already in place. It's just how much more labor intensive it is to scale up from that. So maybe they need fewer people to figure out how they want to allocate capital, how they want to run a business. To the extent we're talking about detecting issues with machinery, you can do that with a building, right? Like preemptive maintenance is an opportunity set for people to get ahead of a generator failing, or an elevator issue, or anything along those lines in terms of how things are being utilized and to try to monetize that more efficiently.

Jennifer Martin

Both of you are sharing that productivity gains will be coming in your sectors, really just from cost optimization and just doing things better, which is really, really important. Finally, could you give us your outlook on AI and its potential longer-term impact on your respective sectors?

Gregg Korondi

Sure, I can start, so let me let me start with a good and a bad. Right. Let's do good cop, bad cop. So, automation in a warehouse is a really interesting dynamic. So, there's been people that have been at the forefront of this for a long time, by putting more robots in the places of people in certain locations.

So, what does that mean for real estate? When you get down to how a distribution facility is utilized, the costs that go into that are frequently equivalent to or less than the costs of automation. So, if you invest heavily in a last mile distribution facility, in an infill location and you build out a huge logistics network that's enabled by AI, you are able to take cost out from other parts of your cost structure and pay a higher rent.

So, the nice part about that is those facilities, they have the improvements and to the extent that any of the improvements are reusable, it provides a lot more runway for people to grow rate over a longer period of time. So, that's, that’s the plus side. Maybe the less plus side is, for a long time, the migratory trends have been to the Southeast and we've seen a lot of companies kind of relocating back-office solutions. In theory, there's usually a big accounting department, there's a big legal department and there's call centers that all service different real estate property types, those jobs seem like they're at risk.

So, if you’re building suburban office buildings in the Sunbelt, are those still in need as office buildings? If you don’t need to hire more people, if you can do more with fewer people. So, the implications around office, particularly in markets where there’s a high substitute-ability of AI technology, is something that I'm thinking a lot about.

Rick de los Reyes

I think one of the points I'd like to make about natural resources, you know, having followed the natural resources sector for a long time, we get a pretty bad rap for being bad for the environment. And I do want to point out that I think AI can actually be a solution here, and not part of the problem. You know, when I tell people what the power demand is going to be for data centers and AI, the automatic sort of reaction is how terrible it's going to be for decarbonization and for the environment.

But the reality is, I think that there are some pluses here that people are overlooking. You know, one is that it takes energy to make energy. Right. And so, all those efficiencies that I talked about before in terms of being able to increase the efficiency in extracting raw materials, all of that's actually going to be positive towards decarbonization.

Talking about using an SMR for steelmaking, you know, steelmaking is a very polluting industry. But if you can do it with an SMR, and be, you know, recycling steel through an electric arc furnace, then that's extremely efficient and very good towards decarbonization.

You know, the second thing I would say is I think all this data center demand could really stimulate more investment in renewable sources of energy. Right. And so certainly the hyperscalers, they're going to want renewables to be part of the solution here. And so, wind and solar are even though, like I said, they can't be 100% of the solution because we need to have reliable baseload power, they're definitely going to be part of the solution and and and those who are making investments in this are going to want them to be part of the solution.

So, I think it could really spur more investment in renewables. And I think it's going to speed up the use of promising technologies, like the SMRs that we spoke about earlier. I think we will see that come on more quickly as we realize that we need to have more renewable and less carbon intensive sources of energy. And so, I would look at this as a real opportunity for, you know, our industries to become more environmentally friendly.

Gregg Korondi

I think that's a great point for real estate, too. If they're thinking about renewables in the sense of monetizing their rooftops, I mean, that's been an ongoing effort. But really thinking about someone who owns a lot of warehouses globally, adding 1.4 gigawatts overall to the capacity to the power network is something that, if we can do that across multiple property types, it can be part of the solution instead of the problem.

Rick de los Reyes

It takes, it takes the investment in technology and it's going to take the investment in the grid because, think about the change that that requires, going from a grid that was just a one-way, a one-way pipe from the power plant to your apartment building. Now, all of a sudden, that apartment building wants to have solar panels on the roof, and it wants to be able to generate electricity and potentially sell that electricity back to the grid.

Right. That requires a two-way pipe that used to just be a one-way system. There's a lot of investment that's going to have to happen here. This is a good thing, and it's a good thing that this is happening, and it's going to need to happen. And I think AI and data centers are going to help drive it.

Jennifer Martin

Well, thank you Rick and Gregg, for your time today in helping us to understand the importance of AI, as it relates to the basic resources and real estate sectors, respectively.

Gregg Korondi

Thank you, Jennifer. Thanks for having me.

Rick de los Reyes

Yeah, thanks, Jennifer. It was great to be here.

Jennifer Martin

It seems we are only in the early stages of potentially significant longer-term impacts on each sector, as AI continues to evolve and is more widely adopted. That brings an end to today's discussion. Perhaps the three key takeaways from today's podcast can be summarized accordingly.

First, AI is already being widely applied within both the basic resources, and real estate sectors to tangible effect. From optimizing costs to improving health and safety, to poring through large datasets to enable better decision making. Secondly, the vast energy needed to facilitate the ever increasing, and more sophisticated, adoption of AI has major long-term implications for both sectors. From securing ongoing supply, to finding key real estate to build crucial data centers. Finally, given the demand for energy, issues around sustainability will be an important, even pivotal, long-term influence on the outlook for both sectors.

So, listeners, watch this space. We certainly will be. Thank you to everyone for listening to today's episode of The Angle podcast. If you would like more information on AI, you can find it on the T. Rowe Price website. And, as always, if you like the podcast, please rate us and subscribe, wherever you get your podcasts.

Disclaimers

This podcast episode was recorded in April 2024.

This podcast is for general information and educational purposes only, and outside the United States is intended for investment professional use only.

It does not constitute a distribution, offer, invitation, recommendation, or solicitation to sell or buy any securities in any jurisdiction, or to conduct any particular investment activity.

This podcast does not provide investment advice or recommendations, nor is it intended to serve as the primary basis for an investment decision. Prospective investors are recommended to seek independent legal, financial and tax advice before making any investment decision.

Commodities are subject to increased risks such as higher price volatility, geopolitical, and other risks.

Changes in the tax laws, overbuilding, environmental issues, the quality of property management in the case of real estate investment trusts (REITs), and other factors could impact the real estate industry.

Investing in technology stocks entails specific risks, including the potential for wide variations in performance and usually wide price swings, up and down.

International investments can be riskier than U.S. investments due to the adverse effects of currency exchange rates, differences in market structure and liquidity, as well as specific country, regional, and economic developments.

There is no guarantee that any forecasts made will come to pass.

The acronym GFC stands for global financial crisis. 

The acronym GPU stands for graphics processing unit.

The Marcellus Formation is the largest U.S. shale-sourced natural gas field.

The Permian Basin is the largest U.S. petroleum-producing basin.

The Haynesville Formation is a major U.S. shale-sourced natural gas field.

The views contained herein are those of the speakers as of the date of the recording and are subject to change without notice. These views may differ from those of other T. Rowe Price companies and/or associates. Information is based upon sources we consider to be reliable; we do not, however, guarantee accuracy.

This podcast is copyright 2024 by T. Rowe Price.

_{ID0006881
202405 - 3502269}