The Dividend Cafe - Data Center Drama

Episode Date: May 22, 2026

Today's Post - https://bahnsen.co/3R9QgGV In this Friday Dividend Cafe, David Bahnsen explains why data centers have become a major economic story, tracing their evolution from 1990s CPU-based server facilities to 2010s cloud-driven hyperscale warehouses and today’s AI-focused GPU centers that require far more power, cooling, and infrastructure. He argues data center construction and related spending may have accounted for roughly 80% of last year’s GDP growth, even as other real estate and industrial activity has been muted, drawing an analogy to the shale/fracking boom. Bahnsen supports data centers and future productivity potential but opposes federal efforts to override local zoning, warns against cronyism, emphasizes the need for a stronger public relations case, and highlights investment implications in adjacent areas like power, water, natural gas, and pipelines. 00:00 Welcome and Setup 00:52 Why Data Centers Matter 01:43 Three Eras of Data Centers 03:51 AI Shift to GPUs 05:42 Data Centers Driving GDP 08:29 Future Productivity Payoff 09:32 What Growth Is Missing 10:12 Fracking Analogy and Backlash 12:15 Localism Versus Federal Override 14:57 PR Playbook Five Points 17:23 Investing Wisely in the Theme 19:35 Wrap Up and Disclosures Links mentioned in this episode: DividendCafe.com TheBahnsenGroup.com

Welcome to the Dividend Cafe, weekly market commentary focused on dividends in your portfolio and dividends in your understanding of economic life. Hello and welcome to the weekly Friday edition Dividend Cafe. I'm your host, David Bonson. I am back in New York City after a week in Miami, Florida, where I very much enjoyed the annual High Tower Advisors Summit. And we had some wonderful meetings and wonderful discussion. there, but got back in New York City last night, and this morning, Friday itself got to write a Dividendon Cafe about today's subject. This subject is somewhat new here in Dividing Cafe. I mean, certainly adjacent to the subject I've written about many times, but the subject itself,

which is data centers and what in the world all the fuss is about as it pertains to a sort of national conversation around data center that's taken over me. many categories of news from markets and financial media, but all the way into a lot of political, local news, all sorts of ramifications, technology itself, obviously. The way it all is a vital integrated part of the AI story. This is not something that is new to many of you as a topic. And it's certainly, as I mentioned, not new in the news cycle. But the way Dividy Cafe is going to cover it today, I think you'll find to be somewhat novel. as is often the case, there is nothing new under the sun.

We talk about topics sometimes as if they're brand new when they're not. And in this case, in fairness, there is some definite newness to discussion of data center, but we need to go back a little bit in time because data centers were not invented with Chatt GPT three or four years ago. I think it became a pretty national and prominent part of commercial real estate life in the United States over 30 years ago in the 1990s, data centers certainly are a major part, I mean, the physical facilities that house servers and that really made the internet possible, even in the early stages of just simply processing email communications.

Broadly speaking, a factory that rather than having widgets or parts has digital services is not new. The data that has been stored and processed to make the internet possible was really a part of data centers going back into the 1990s. Now, a lot of it was smaller businesses in particular were in a closet. IT centers that had a stack and a rack and servers right there. And I want to make one point now that I'll kind of elaborate on in a moment. These warehouses were filled with what's called CPU, central processing units. And that's going to be a distinction from where we are now for a reason get to. But then I think the second phase, which became a very big deal, and now these physical facilities started to get a lot larger, and people started to see them a lot more in various parts of suburbia

or out in more rural areas or whatever the case may be. But it's the kind of 2010's data centers are different than the 90s and definitely different than the AI story we're getting to in a moment. But what I'm really referring to is that which made the cloud possible. And now there was a need for a lot more storage. First of all, smartphones took off, but the cloud computing elements really got hypers involved. And you did have massive facilities that at this point were housing tens of thousands of servers managing the data, the traffic of cloud, of social media, and of streaming. And then, of course, this kind of third iteration data center is the one we're going to talk about right now. And first and foremost, I should point out that they're largely housing something a bit different

and CPUs, and that is GPUs, which is graphic processing units. And I'm going to explain why that matters here in a second. But the AI infrastructure, the need for this source of power for AI generative models, your chat GPTs, your quads, your Gemini's, they bring in, first of all, significantly more processing of computation, which requires a lot more space and a lot more infrastructure. But I don't want to get into the tech side of it because, first of all, it boars me to tears. And I got to think it's going to bore a lot of U to tears. But I think that why I bring this up in terms of just the massive increase exponentially

so in computations being done. So therefore, the requirement for a lot higher quantity of these centers, which themselves, each one has a higher requirement of physical space. But that most importantly, being basically a home for GPU. that take about 10 times more power than CPUs do. And it's oftentimes more than 10 times. And so then now you have a much higher energy demand and much larger facilities that require a lot more cooling.

So there's a water and energy and power component as well. So they're not glorified server closets like the 1990s, but they're also not just mere warehouses like the 2010s. These are digital libraries, and they are of an industrial and technological, sophistication, size, scale that is very unique. Well, why is an economist finance investment property, like Dividing Cafe,

talking about this granularity of technology that is data center? And the reason is this has become just a massive economic story. without boring you with too much data. The reality is that you could make an argument that data center construction and some of the adjacent components to it represent 100% of net economic growth last year. I think it's closer to 80%.

There's some cases for it to be smaller because the reason why it's gray is because it depends on how you're measuring certain things. Harvard economist Jason Furman, who was in the Obama administration in which there's an awful, lot of things that I disagree with him about. He is a Neo-Kinsey into the core, and I am, shall we say, not. And yet, I know him to be a very straight shooter and honest broker, even when I disagree

with him. He calculated investment in technology to be about 92% of GDP growth last year, but that's including more than just AI infrastructure spend. It's including equipment and software, and so there's some murkiness there. Renaissance macro research estimated the data center buildouts were higher dollar contribution to GDP than all consumer spending was. Now, these things are weighted differently percentage wise, but the dollar contribution, that's unbelievable. Okay. So what we know is it's a big, big deal, regardless of how you're exactly measuring it. Total construction spending last year was down 1.4% and data center construction was up 30%. So think of how much construction would have been down without data center. And by the way, that's construction,

meaning there's shovel in the ground. There's money for approved projects. The amount that is committed or being pursued or targeted because it is not yet entitled, zoned, approved, et cetera, is exponentially more. We're talking about actual dollars spent. The data center subcategory of construction was four times higher last year. than 2021. So we're seeing something that's significant industrial production for IT equipment is up 80% since the beginning of the decade. The rest of industrial production dead flat. So I'm in the camp. S&P Global did a great study that suggests data center construction and adjacent is about 80% of last year's GDP growth. And I think that they make their argument cogently. So this is a big deal for the

entire state of economic growth. The question, though, is more than just what it represents now in economic growth, because these hyperscalers are not spending this kind of money. There is not this sort of investment going into data center if there isn't some belief that it contributes to a lot more economic growth later. At some point, you have to get some productivity on the other end of it, what the data centers are building and facilitating. I've talked about this before with the overall AI dominoes. The language model, have to then be essentially transferred into some sort of product that is used by businesses at an enterprise level or consumers to drive productivity and drive quality of life and all these things.

And there's lots of arguments that's going to happen in spades. How it's all economically measured? We don't exactly know. But what's on the line here for data center investment is not just merely what the investment is in the here and now, but what it's supposed to be creating for the future as, well. This is a conversation about present economic growth and future economic growth. And it kind of helps to mask something, a sort of quiet part of this that needs to be said, which is, if this is such a huge percentage of economic growth right now, what is not taking

place in the economy? Well, very muted, multifamily development, office construction has gone negative, overall commercial real estate, even factory construction that's non-data-centered are oriented has gone completely flat. Across the whole industrial cycle, there is just this really significant contribution from data center and then either negative or flattening across everything else. And so the economy has kind of gone all in on a particular play. The analogy I become fond of using, first of all, I'll give a hat tip to my friends at Chutegas Research because this was a sort of articulated comparison that I first got

from them is in comparing into the fracking, the shale revolution. if you will, of 2010, 11, 12, whereby significant capital expenditures were going into hydraulic fracturing and horizontal drilling, but almost nothing else post-financial crisis was happening there. So he had a disproportionate contribution from a single sector, single investment, and that's a very good thing, but it was masking that there was a lot of muted economic activity and other sectors of the economy, and it's really quite similar. But where Straticus research was making an analogy is in the other part that I want to point out now is that this was a very disproportionately positive part of economic growth with fracking 15 years ago and with AI data center now. But it was very politically unpopular.

It was very culturally problematic. You had a lot of people against it even though it was like the big thing going on in the economy. Then it was much more partisan. It was divided up among the red state, blue state thing, and there was definitely an administration at the time. I think the Obama administration did not get enough credit for being reasonably supportive of what was going on. But a lot of the reason they didn't is they didn't want to talk about it because their own base didn't really like it a lot. And that's not quite the case here where there is a significant concern and cultural stigma around this major catalyst of economic activity with AI. But I don't think it's as partisan.

I think it's either bipartisan or nonpartisan, depending on how you look at it. And so there's a great analogy there in the data center moment now to fracking men of some questions about public sentiment and popularity in both political and cultural context around this thing that is very economically catalytic to current growth. And in fact, we're dependent upon it in the very short term. My own position is like so many things these days requires nuance and it gets lost in a black whole in nuance where I would say to you, I'm overwhelmingly for data center and I'm overwhelmingly against the federal government telling cities, counties, local places that they have to bypass

their own ordinances to get these things done. I think that there should be the right separation of powers. There should be 10th Amendment federal localism, federalist localist approach. And if a particular municipality doesn't want a data center in their backyard, they don't have to have one, and then there's trade-offs to that, and they will live with the trade-offs, but I do not support Washington, D.C. running roughshod. So on one hand, I think a lot of municipalities are wise to say we're concerned about the energy consumption, we're concerned about water needs. I mentioned the cooling factors earlier with GPUs versus CPUs, and I think that there are some that are just rank nimbies that are saying, I like not looking at things I don't

like looking at so I don't want a data center 22 miles from my house that I might have to drive by. Those things are a bit different, but the point being, there are zoning laws and entitlement ordinances at play that I don't think should be bypassed. If they're not being bypassed for a local pub or a church or a Costco, then I'm sorry. I don't think Silicon Valley billionaire should get them bypassed by Washington, D.C., no less. So all at once, I can be pro-data-center and pro-growth investment for the future for risk takers. I think the cost of the risk taking should be borne by the risk taker. And I think it mostly is, by the way, but I do not support crony deals and I do not support violation of the 10th Amendment. However, I do support data center and growth and investment and so

forth. So if somehow people can wrap their arms around the fact that both those things can be true at hand, maybe you can follow the consistency of what I'm trying to suggest. But what's going on now, township starting to pass ordinances blocking all data centers, even where about 25% of hyper-scaler computing comes from, power comes from. This data center corridor in Northern Virginia, Loudoun County is starting to push back. Certain states have floored with bills. Maine actually passed one at the legislator and then the governor vetoed it. So this thing is not done. I mean, the political toxicity around us at either a city county or state level is significant.

And I would suggest that's okay, even though if I were in some of the localities, I might be supportive of the blockers in one and not supportive of the blockers in another, depending on those local circumstances, okay? But where do we go from here? Well, what's ironic is you absolutely have to have these data centers built to have this whole thing come together and for the computing power that is required and the electricity generation needed to fund these GPUs. If this whole story is going to happen, you need more.

But I'll tell you, the CEOs of these companies don't act like they want to win the PR war very much. And they're going to have to. But in the meantime, I'm perfectly fine to say that the tradeoffs that exist and some people wanting data centers and others not, it's okay. And everyone can live with those tradeoffs. Let me read it to you the way I put it in Dividy Cafe here. I don't think making the case out of fear and panic is going to work. I don't think scaring people that China is going to beat us is going to work. But I also don't think that not making the case is going to work.

We really need data centers and they're going to benefit your lives so much, but we don't care if we have your approval or not because we're just going to ask the White House to override your city council. I don't think that's the right play to win friends and influence people. The need of the hour is a massive public relations effort that would be economic, political, and cultural, but done in the spirit of the social contract of our economy. And it has got to be rooted to, and I'm going to give you five things here. A, the economic reality, the data center AI CAPEX is vital in the present economic moment.

That's a pretty easy case to make. So make the case. Number two, the economic reality that if AI is going to deliver the future promise productivity growth that people have promised, it's vital for a future economic moment. Go make that case. Number three, the local sensitivity. We have zoning laws. We have community processes for a reason for good or for bad.

So honor those processes the way other businesses do and do not seek to bypass them just for billionaire Silicon Valley folks. Number four, a disdain for cronyism has got to be front and center where water, power, other costs are socialized, but the profits are privatized. That has got to be rejected. Number five, the realization of the energy needed to power the construction and investment requires more natural gas. and more pipelines to get the natural gas there. If you can't make that case, and if you don't understand that, you don't understand the whole data center story. I don't think there's enough totality in this entire narrative.

I'm not against building more data centers. In fact, I'd argue right now it's the biggest thing we have going for the economy. The data is rather clear. But I assure you that not all of these data centers are created equal, not all the data center sponsors are created equal, not all the projects are. Much like the pipeline analogies 15 years ago when I talked about fracking Shale revolution, there's different counterparties, there's different funding models, there's different contracts.

There will be winners and there will be losers. And that was so, so true in the Shale Revolution and it is so, so true in Data Center. When it comes to direct data center investment, be discriminating. Sponsor quality is paramount. And I'm telling you right now, you're going to see how right I'm about this. And there's people that do not take this seriously. They're going to be wrong and they're going to learn it the hard way. I don't think there's any basis for rejecting societal norms and constitutional structures to accommodate commercial actors.

I think you can let localism play out and let different communities experience those tradeoffs the way they want to. And that will lead to some that are very pro-data center, some that are not and some that have different particulars. And I think that's fine. The adjacence to this story for investment opportunity is where I would focus. Not necessarily the story itself, power generation, utility output, water, natural gas, natural gas, pipelines. There's a lot more to this from an investment standpoint than just the mere data centers. But in the end, like AI, data center is not an end. It is a means to an end.

And the data centers have to create computing power. so the computing power can fuel AI tools so the AI tools can drive productivity. And those dominoes have a lot of risks and rewards within them, stuff that can go wrong, stuff that can go right. That is the investment story here. That is the broader point I'd make around the data center particulars as we fight through this as if it were just a NIMBY story.

There's a lot going on. I hope this is helpful for your thoughts. when you think about this from an investor standpoint. And I welcome your questions anytime. I love the topic. A lot more to say. In the weeks, month,

years ahead. In the meantime, thank you for listening. Thank you for reading. Thank you for watching the Dividendon Cafe. Monday is Memorial Day, so we'll have no dividend cafe on Monday, but I'll be with you next week.

I'm in different cities in Texas each day next week, speaking and meeting with clients. And then I'm at the Reagan Library. Next Friday, as part of the Reagan Foundation's annual economic forum or I'll be speaking with Secretary Pescent and Jamie Diamond, amongst other luminaries. Looking forward to that and I believe I'll be recording the Dividendon Cafe from the Reagan Library next Friday. In the meantime, happy Memorial Day weekend. Take care. The Bonson Group is a group of investment professionals registered with Hightower Securities LLC,

member FINRA and SIPC, and with Hightower Advisors LLC, a registered investment advisor with the SEC. securities are offered through Hightower Securities LLC. Advisory services are offered through Hightower Advisors, LLC. This is not an offer to buy or sell securities. No investment process is free risk. There's no guarantee that the investment process or investment opportunities referenced Tyrion will be profitable. Past performance is not indicative of current or future performance and is not a guarantee.

The investment opportunities, reference Tieran, may not be suitable for all investors. All data and information referenced herein are from sources believed to be reliable. Any opinions, news, research, analyses, prices, or other information contained in this research is provided as general market commentary and does not constitute investment advice. The Bonsor Group in Hightower shall not in any way be liable for claims and make no express or implied representations or warranties as to the accuracy or completeness of the data and other information, or for statements or errors contained in or omissions from the obtained data and information referenced here in. The data and information

are provided as of the date reference, such data and information are subject to change without notice. This document was created for informational purposes only, that opinions expressed, are solely those of the Bonson Group and do not represent those of High Tower Advisors LLC or any of its affiliates. High Tower advisors do not provide tax or legal advice. This material was not intended or written to be used or presented to any entity as tax advice or tax information. Tax laws vary based on the client's individual circumstances and can change at any time without notice. Clients are urged to consult their tax or legal advisor for any related questions.