[00:00] Bryan Hyde: Fall flavors are waiting for you at Whole Foods Market. Cozy picks all through the store, like Maple Leaf Cookies from 365 Brand. Swing by the bakery for Pumpkin Cheesecake, here for a limited time, and Pumpkin Turmeric Bread, freshly baked every single day. Sip the season with 365 Brand Cinnamon Spice Coffee and Pumpkin Spice Creamer. While you still can, spice up fall at Whole Foods Market. [00:27] Domenico Ferraro: Today I'm joined by Domenico Ferraro. He is a professor at Arizona State University, the W.P. Carey School of Business. And this is only scratching the surface though. And Domenico, take a moment if you would. First of all, welcome to the show. Take a moment to tell us a little bit about yourself and, and who you are and what you do. [00:46] Speaker C: Hi, Brad. It's great to be here. Thank you for having me. As you said, I'm a professor at the W.P. Carey School of Business at Arizona State University. And I spend most of my professional career trying to understand economic growth and the functioning of the labor market, mostly the US labor market. But over the last year, I got interested in the question of data centers. And as you know, of course, there is a lot of debate about it. So I like to paraphrase a famous economist, Robert Lucas, when he say, if you, when you start thinking about the data centers and AI, you can stop doing it. Okay, so here we are. I started one year ago, I got interested in it, and I wrote the 3 reports for the Godwater Institute. And I'm really glad we can talk about them today here. [01:41] Domenico Ferraro: Let's— before we dive into those reports, like you, I hear a lot of people talking about data centers and sometimes connect— they connect it with AI, but there's so much talk about these data centers and I hear a lot of fear. In fact, I hear so much fear, I'm not really sure what to believe sometimes. There's some really stark predictions being made. Could you just kind of set the stage for us about, let's just kind of give the thumbnail sketch of when we're talking about data centers and AI, let's have kind of a working definition of what we're talking about. Then we can get to whether there's reason to be concerned or whether these are being overblown, who's benefiting, and so forth. [02:22] Speaker C: Look, I'm very skeptical of doomsday scenarios, but I also am aware that historically this type of reaction has been the norm. At the outset of technological revolution, there is always a concern that this technology, while could have lots of value, It could have some disruptions in the labor market. And how society are going to deal with these disruptions can create further problems, which instead of creating an environment which is conducive to the appropriate adoption technology, actually it gets to social conflict. So in that sense, concerns are valid. Because historically it remains true that technology can create winners and losers. However, my stand is let's think coherently about what is perhaps the better way ahead. Okay. So we have a technology like AI and the data centers that power them, that there is no doubt is going to have enormous value. In the future, I think we have only seen part of this value today, but the most of the value, how is going to change our life, every facets of our life, I think is going to be five or ten years from now. Okay, so there is enormous value. Now basic economics tells us that if there is an asset, if there is value, most likely in order for us to grab this value. We need to incur a cost. Okay, now what is the cost? The cost is we need to buy equipment, we need to retrain ourselves, and we need electricity because ultimately computers, service servers, cloud computing, they need energy, they need to be powered. And that's where the data centers comes in because data center is the physical infrastructure that enables artificial intelligence, but not only artificial intelligence, everything that has to do with computing power needs electricity, needs to be energy. And that's where data centers come in. And then, of course, as I said, if there is a cost, then the questions that I think we can ask is that who is incurring this cost? Who are the people that are going to be facing the risk of displacement in the labor market, if any? Who are the people that are going to pay this energy cost, so on and so forth? So, in that sense, concerns are valid in a sense that when we talk about cost, most likely, they're going to be people that are going to face a larger part of this cost. But doomsday scenario, I think the They, instead of thinking in a coherent way about the challenges and opportunities, they only focus, they only zoom in on the challenges, forgetting these enormous values ahead of us. So that's the way I think about it. [05:41] Domenico Ferraro: I think that's a really healthy way to approach it. And for those people who are really prone to look at, well, here's the bad that could happen, I like to imagine what if you could bring someone from 1,000 years ago and just give them a glimpse of what we have. If you took them into a supermarket, they would probably think they had died and gone to heaven, you know, based on what they're seeing there. And so we have to be open to— there could be some very good things. And I'm with you. I think this is opening up some great possibilities, but we have to navigate a few tricky areas first. Now, you've written 3 reports for the Goldwater Institute. Let's touch on those reports. And I'm going to— I'm going to leave it up to you to just kind of give us a summary. We don't have time to go into great depth, but I would love to hear what you have found and what you have written on. [06:27] Speaker C: Yeah. So these 3 reports, of course, they're all related, but each one of them focuses on a specific aspect of what I think are the criticalities of artificial intelligence computing and data centers. They all point to the value that data centers are likely to bring to society, but they also touch upon in a, I would say, serious way about the cost associated with data centers and improved computing in the future. So one of the reports really zooms in on the value of computing artificial intelligence. So what is the value? I think most commentators really focus on the value that AI and associated data centers powering them are going to bring to the labor market. Okay. And of course, I understand why, because the analogy, think about computers, for instance, right? Computers had effects on the labor market. [07:42] Speaker D: Why? [07:43] Speaker C: Because computers made most workers, not all workers, better at doing what they do. Okay, so this is improvements in productivity, wages go up, we are all good. Okay, now this is only part of the value though. Now most of the debate is about how this new technology is going to affect the marketplace, how it's gonna affect the production of goods and services that we buy every day. Now, this makes sense because most commentators think about a technology that now we have grown accustomed to, which are computers, right? Computers was a major transformation of the way we work. Okay. Now, part of this report on the value AI also points out that there is also value At home, home production. Okay. Some of our activity, actually a non-trivial part of our time in a day, is spent creating goods and services that are not sold in the market, but they're sold in the household. Okay. For instance, just a concrete example. First of all, we are doing a podcast and we are using a computer. And most likely we are using the power coming from a server that is sitting in some data centers. Okay, so here we go with this one example. The second example is that suppose I like eating pasta. Okay. Unfortunately, I don't know how to cook pasta. Now, if a few years ago I would have asked somebody who is an expert in cooking pasta, I would ask for a recipe, perhaps made a phone call to make sure that I would cook the pasta the right way. Right now, I can ask ChatGPT, I can ask Claude, I can ask Gemini. I want to cook this type of pasta. Please give me some recipe and some suggestion of what to do and what not to do. It's more like having a grandmother, a digital grandmother. Okay, and that's very valuable. It is very valuable if you think about it, because To the extent that I really like eating pasta, as I do, for me, having this resource that is easily accessible on my iPhone or my computer creates a lot of value for me. [10:19] Bryan Hyde: Okay? [10:20] Speaker C: So the punchline of the first report then is really to point out that artificial intelligence has the power to transform our lives. in the workplace and at home. Okay. Now, yes, of course, if you think about the workplace, there are valid concerns about whether some people are going to lose their job, whether there's going to be a retransform— restructuring of entire sectors of the economy, which requires a few people to be unemployed for a while. Those are all valid concerns. But my point is that if you look at the data now, if you look at the evidence now, there seems to be no major disruption in the labor market. Okay? Someone could say, but we have not seen it all. And that's fair. That's a valid concern. Perhaps this disruption, I doubt disruption is going to be 5 years, 10 years from now. But if I look at the data now, I don't see any of that. I know that the value of artificial intelligence And AI is big now for me because I use it every day. I know that the value at home is large because I have my digital grandmother, a digital grandfather that can help me, right? [11:41] Domenico Ferraro: Yep. [11:41] Speaker C: So my premise is that, okay, there is a value right now. We can quibble about how big it is, but my understanding through the use of it seems to be This is a transformative technology. [11:58] Bryan Hyde: Okay. [11:58] Speaker C: And I even think that a bigger chunk of this value is ahead of us. Okay. So that's the premise. Okay. The premise is there is value there. Now, economics tells me that there is no— nothing is for free. We would all hope to live in a world where there is a great technology with enormous value, but there is no cost. Okay, this is not the real world. Okay, so the real world is one in which if there is an investment opportunity, if there is a technology that has the transformative power as artificial intelligence and the data centers associated with, then we need to do something in order to grab those benefits, in order to reap those, that value ahead of us. Okay. And we call it cost. Okay. Now, the other 2 reports are the tackles the cost side of the data center problem or the economics of data centers. Okay. So what are the costs? Now, I think there is no doubt that we need to expand the capacity of the US economy to produce electricity. So we used to live in a world where the system overall was able to generate enough electricity to all activities that we daily engage with. There are signs of strained grids and regional electricity markets that tell us that we are about to enter in a new era, in a new regime. The new regime is one in which the demand for electricity is substantially larger than it used to be. Part of it is coming from data centers. [14:06] Speaker D: Okay. [14:07] Speaker C: But again, data center is just a conduit Is the way we power improved computing, is the way we power our digital grandmother and grandfathers. Okay, so that's the premise. Now, how you expand capacity? You need investment. There is no way around it, right? There is no magic wand or no trick. There is no Tactics, if you wish, right? We need the strategy, right? We need to expand the ability of the U.S. to produce energy, produce electricity at fair prices. Okay, this requires large investment in electricity generation. Now, how we do that? How we achieve that? That's not easy. Why is it not easy? Because investment in electricity requires big fixed cost upfront, right? You need to build new generation from scratch, new transmission lines, new distribution lines. And the investment is today, it's going to take time to produce this new generation. So if you think about it, think about someone that's trying to decide whether to invest or not in new generation, you say, well, I invest lots of money today, but the demand for new electricity is going to come 5 years from now or 10 years from now. Well, should I really do it? Then you start some calculations and you say, well, what is the interest rate now? Now, we live in a period of higher than average interest rate. So, the cost of capital is high. And then you start also asking yourself, well, is the regular— are the regulations in place clear enough of what's going to happen 5 or 10 years from now? To the extent that I perceive some uncertainty about regulations, That even further reduces my willingness to invest. In that sense, investment in electricity more than in other sectors requires a lot of coordination among several stakeholders, requires coordination between the data centers that are demanding electricity, coordination with those that are in charge of investment and producing new electricity, new generation. And then there are the regulators that participate in this relationship in setting the rules of the game. Now, you can see that this becomes complicated fairly— these calculations become complicated fairly easily. So in that sense, one would need a lot of coordination among stakeholders to create an environment that is conducive to large investment, which are needed to expand the capacity. Yeah. [17:34] Domenico Ferraro: So we're down to our last few moments here in the segment, but are there any must-have policy objectives that need to be in consideration? I know I've just asked a question that you could probably talk about for a very long time, but When it comes to moving forward and doing so in an informed and productive way, what are some of the top policy objectives that have to be considered with responsible AI deployment? [18:00] Speaker C: So that is a great question and it's also a difficult one. So to me, I see clear principles that one should keep in mind when thinking about safe deployment of data centers, where I'm using safe in the following sense. There is a risk that large load, demand load of electricity come in. The system in the short run does not have enough capacity to meet this increased demand. And this implies that the price of electricity must go up. There is no other way around. Okay. Now, there is a risk that the increase in the cost coming from this increased demand is going to be sustained also by those consumers that demand very little energy. Okay. So the risk is that this increased demand is going to congest the grid. The price of electricity has to go up to equate the demand and supply. But then the question becomes, what are the distributional consequences of this increase in price? Now, there are different ways to go about it. One would say, well, Domenico, you told me that the price has to go up. The price is going up. But the risk that I envision here is that the price is going to go up for everybody. And this is not good economics because good economics implies that we should think of willingness to pay. If there is a big data center that comes in that needs a large load, then this also implies that the willingness to pay of this data center arguably is larger, if not much larger, than the willingness to pay of Domenico, who is a small consumer of electricity. Therefore, an appropriate pricing system should take this in account and price electricity based on willingness to pay. Now, some electricity markets don't fully take this in account. Some others are completely predicated on a different logic where prices are set for everybody in the same way. So in that sense, the principle that I want to convey here today is that one, if you believe that there is enormous value, investment must come. In order to have this investment, some coordination among stakeholders is needed, including giving the ability to data centers to arrange their own electricity supply and electricity generation with private contracting, with independent producer of electricity. 3, be aware that the way we get to increased capacity is— can be tricky because there could be small consumers of electricity that end up bearing the cost of this increased demand. And that is not what Good economics would tell you to do. Willingness to pay is key here. [21:35] Domenico Ferraro: All right, that's the note where we're going to go ahead and pump the brakes. I hope we get a chance to pick up this conversation again in the future. Again, we're talking with Domenico Ferraro. He is an associate professor in the Department of Economics at Arizona State University's W.P. Carey School of Business. And thank you so much for joining us today. [22:03] Speaker D: This is Sargon with Sargon Law Group and your Legal Minute. Kids are back in school, which means our roads are changing too. There are new teen drivers behind the wheel, school buses making frequent stops, busier crosswalks, and families learning new routes. Slow down in school zones, put your phone away, and give yourself a little extra time to react. little extra time. A few extra minutes can help prevent a serious accident. Phoenix, if you're hurt, call Sargon Law Group or visit callsargon.com.