I assess the state of the data center and the AI infrastructure buildout given the big election sweep by the Democrats. Many Democrats were running on moratoriums and ban themes.

In a video analysis, how election-driven policy changes may affect U.S. data center and AI infrastructure growth are assessed.

Read more

Power is where the AI buildout gets real. Many of the headlines about AI are snippets. The real constraint on compute is power, and power is what solves it. In the November D CEO energy edition, I tell the tale of that story through two features: one is a profile of Vistra and its CEO Jim Burke, who received the Legacy Award at a recent energy awards event. An accompanying trends piece features Applied Digital CEO Wes Cummins. Both are Dallas-based chief executives solving power challenges in the AI buildout.

Myths are circulating in the ether that energy demand is explosive. Burke notes that electricity demand is growing roughly 2 to 3% a year. Another overstated statistic is the widely cited 400-plus-gigawatt Texas queue for large loads, of which is 90% data centers. This absolutely will not be built. Burke sees more like 12 to 15 gigawatts through 2030, possibly 15-20. This may happen beyond 2030, given the Texas audit and therefore potentially slower pacings depending on the election outcome. Interconnection queues need to be sorted to identify real projects, and both the PJM and ERCOT grids are working towards this in their own ways.

According to my work, Vistra saw the need for “dispatchable” or firm baseload early. In 2018, when the conversation was all about renewables, Vistra bet that firm, dispatchable base-load power would be needed. It bought undervalued gas plants (adding almost 20 gigawatts with the Dynegy acquisition) and invested in nuclear well before Amazon’s Susquehanna deal woke up the power industry. It’s market cap grew from roughly $10 billion in 2017 to about $47 billion at the time of print. Burke, at the company 25 years, has paralleled the Texas market’s competitive and deregulated landscape also for 25 years.

There are many paths for greater efficiency in grids ahead. Vistra operates in both PJM and ERCOT, the two largest U.S. markets. Its mantra is to use existing assets more smartly—the grid is only about 35% utilized, according to Burke. Vistra has extended the lives of several nuclear plants around the country. Grids are built for summer peaks, but winter peaks are increasingly a source of grid challenges. Batteries and renewables help, but batteries cannot cover a multi-day winter storm, so firm power remains essential. The goal is to power AI reliably, affordably and efficiently.

Adding large loads can actually lower ratepayer costs, which is an underreported angle. Hyperscalers are less price sensitive than past industrial customers, so grid upgrades are happening on the back of hyperscaler capital spending. These large loads are paying for infrastructure upgrades; they also help spread fixed costs across ratepayers over time with their additions, potentially reducing costs. With Vistra’s generation fleet, Meta and Amazon are paying for nuclear fleet extensions, giving ratepayers carbon-free attributes that would not otherwise exist. This rarely shows up in the headlines.

The Texas market is a proving ground for the AI buildout. It is also the most competitive energy market in the world, with long experience in co-location approaches, for example, pairing a power plant with an industrial asset such as a petrochemical plant or data center. Regarding the idea of flexible loads, an approach needed ahead, Texas’ bitcoin miners curtail their energy usage during grid emergencies or in response to higher price signals. To deliver speed to power, behind-the-meter builds, a more expensive option, are also emerging. Most of these designs, funded by hyperscalers or investors, will ultimately be added to the grid.

In light of the moratoriums proposed, grid-connected assets appear the way to build. Still, one-sized approaches and prescriptions do not suit every case and every locale. The market and the public like the simple answers. They do not exist in such as complex tie-up between energy and digital infrastructure.

The market is still conflating long-lived assets versus an asset-light mindset inherent in the hyperscaler’s former life. Power plants tend to last 30 to 50 years, and digital infrastructure is also long-lived. Big tech is used to asset-light models and is now bumping into heavy infrastructure spending, which helps explain market volatility and a misreading of the play. Infrastructure firms and big finance are entering the market and may be the eventual owners of these assets in the future as well. A Blackstone REIT and hybrid infrastructure solutions characterize the new day. Some of the reactions about debt financing and neocloud constraints circle this dynamic.

Hyperscale capex has been a major support for U.S. GDP. Top hyperscalers will spend over $800 billion in 2026, but thus far 2027 spending looks higher. Peak spending could come around 2028. None of this changes the impact of the AI boom, just where the revenue and costs shift. In the next 3-5 years, a baseline will emerge. In the last year, the different paths of hyperscalers, and the disruption therein, has become clearer.

Oil and gas, utilities and independent power producers like Vistra are integral to the AI buildout. Renewables and nuclear are being reupped in the mix too. The additional energy demand will require an efficiency-first mindset.

The Interview

ANNOUNCER: This is KGVO’s Talkback. Join the conversation at 721-1290 or 1-800-568-5309 with producer Nick Christensen and your host, Dennis Bragg.

DENNIS BRAGG: And good morning, everybody. Dennis and Nick back with you. We’ve got a little bit of sunrise pink out there. That’s kind of pretty; it distracted me for a minute as I’m looking out the window. It was a pretty interesting weekend. Did you have an extra hoodie on this morning?

NICK CHRISTENSEN: It actually wasn’t too bad. The frost was on the windshield a little bit, so I had to wait a second to go out and clear it. But no, it wasn’t too bad. It was a nice weekend.

DENNIS BRAGG: This is the first morning on the drive in where I thought, “I should know where my gloves are.” So we’re going to warm up the airwaves, as we have a great topic: data centers. Bob Seidenschwarz is not in the studio, but he’s out on remote, traveling, and took some time out today to join us. Bob, why don’t you go ahead and introduce our guest this morning?

BOB SEIDENSCHWARZ: Oh, it would be my pleasure. But guys, I’ve got to stick it to you a little bit. I’m looking out the window, and the dolphins are jumping in the ocean.

DENNIS BRAGG: See, just stop that Florida stuff. I don’t wanna hear it.

BOB SEIDENSCHWARZ: Eighty degrees and no humidity. It doesn’t get any better. So you can chew on that for just a minute.

DENNIS BRAGG: Okay. Thank you.

BOB SEIDENSCHWARZ: We’ve got Jennifer Warren, who has been a guest of ours, and we’re very grateful. Again, I want to thank her for taking time out, because this is such a big topic. We’ve had numerous discussions, and I don’t know anybody better at dissecting what the opportunities and challenges are. So let’s get right into this. Jennifer, welcome back. You went through and are going to be publishing a pretty extensive paper. There’s video out on this and discussions. So let’s get started with the AI build-out on the infrastructure side and the data side, and specifically hone in on North Texas leadership. This is an epicenter, very similar to Northern Virginia, as the two largest areas where this boom is really taking place. So welcome back again, Jennifer, and let’s start from that perspective.

JENNIFER WARREN: Okay. So what I can say is the US is really a growth engine of the global economy right now, and it’s largely because of this AI infrastructure build-out, which includes energy infrastructure, computer and IT manufacturing, and semiconductor manufacturing. It’s contributing almost half of our GDP growth, which was 2.7 percent in some of the last figures. It’s actually a little bit higher than what gets understood, because when you back out all the imports, it looks lower. We are importing a lot of chips from Taiwan and South Korea, but a portion of that manufacturing is coming back to the United States, and that’s helping boost growth over time as well.

JENNIFER WARREN: And yes, North Texas has been something of an epicenter, because we have a lot of data center developers on the forefront of building the infrastructure where much of this AI technology will be housed. So you’re getting what’s called an AI factory, if you will, and that’s what a lot of the hyperscalers, which are big tech, are spending on. Manufacturing is also contributing considerably to the Texas economy, and we’ve seen that in the data as well. It’s really starting to show through GDP growth.

BOB SEIDENSCHWARZ: So when we’re talking specifically about North Texas, one of the criticisms of data centers is, “Oh, it’s temporary. Once the facilities are built, and we’re promoting jobs, are those jobs going to stick around?” Can you give examples of where we’re seeing the secondary and third-tier growth from these data centers and the infrastructure build-out? I realize we’re talking about different phases, so if you want to walk us through how these are not one-and-done deals. These have extensive, long-term economic implications for our growth.

JENNIFER WARREN: Well, I can give you two examples. There’s the Stargate Abilene data center, which right now has around 1.2 gigawatts of capacity. It’s going to grow. Microsoft is a tenant, and OpenAI. It’s definitely created a boom in construction, and that’s going to go on for some years. But there are a lot of businesses surrounding that which are growing as well.

Talking about the larger ones. There’s one in Louisiana called Hyperion, which is Meta’s data center, and it’s supposed to be 2.3 gigawatts. Just to note, each gigawatt is roughly a $50 billion investment. It’s a huge amount of money, and I think the hyperscalers are going to be spending roughly $800 billion this year, and right now it’s already looking like it could be higher [for next year]. In Louisiana specifically, Meta is contributing considerably, not only to the tax base. For example, teachers had something like a $50,000 bonus because of how it affected the revenue intake of the area. They’re also going to be funding water projects and building their own infrastructure tied to the data center.

That’s because that state and that area particularly want growth. That’s what you’re seeing with these large-scale campuses: they’re largely going to be locating in communities that want the growth. It’s also the case in the San Antonio area, and West Texas is attracting more large-scale developments. Those are what are called frontier markets. And in fact, with some of these data center bans and moratoriums, Texas picks up business, Montana picks up business, as does Nevada, because of the business climate and the availability of fiber and energy. That’s also a big draw in Louisiana, because there are a lot of natural gas resources.

DENNIS BRAGG: That’s interesting. As we continue on this Monday edition of Talkback with Bob Seidenschwarz and also Jennifer Warren, we’re talking data centers. Jennifer, I want to bring AI into the conversation here. Obviously these data centers are used for more than just AI, but that’s a lot of the emphasis. I’m curious because in the last three or four months, since summer, I keep seeing stories and a debate going on in the tech community: are we all going to go with the frontier models, or are people going to have it on their desktop? I think that’s an interesting conversation.

DENNIS BRAGG: It reminds me of the old mainframe debates in the early eighties. Is it going to be all remote, or on the desktop, or a combination? Since we’re still working out where that lands in the coming years, how is that influencing this push toward data center investment?

JENNIFER WARREN: So, two things. The AI build-out consists of not just data center build-outs; it includes the build-out of energy and infrastructure. It includes a lot of R&D, building these frontier models like Anthropic, OpenAI. Google has its own frontier model, Gemini. Meta now has Muse. There’s just a lot happening in that space, and a significant amount of investment going there. But you’re also seeing more AI applications, and the expectation is that that’s where a lot of the future value could be, as businesses start creating around AI technology, using AI technology. You see a lot of that in big tech. They know how to use AI technology and frontier models, right?

You’re seeing service providers around that also creating new value, finding new offerings to help businesses be more efficient, but also finding new business models. For example, something really interesting that’s happening with the data center. If we just call it digital infrastructure, which really reflects what it is, because data centers evolved from commercial real estate. That’s why there are a lot of these developers in Dallas: we have a lot of commercial real estate activity, and it’s a big part of the Dallas economy. So we have this expertise that’s morphed into digital infrastructure development, which is an industrial infrastructure asset, if you will.

Data centers do have cloud activity, but they are also starting to produce what are called tokens, which is really what the frontier models are cranking out when you’re using AI and it’s doing the inference, and people might be using agents. Usually that’s going to be tech people for now.

Now, as far as the device question goes, I interviewed somebody for an upcoming feature that should hit print at the end of the week and be digital probably next week. He thought that, because of the scale required in these big campuses to house all this IT, there’s going to be a lot happening on device. That can be laptop; that can be mobile. But that means data centers could be sited in a lot of different places, because with fiber networks, you have the ability to do that. In other words, data centers don’t have to be sited where all the activity is. There’s also a lot that needs to be close, because of latency and issues that we can’t go into. You can’t really say anything hard and fast at this point, frankly, because there’s so much disruption and churning.

DENNIS BRAGG: It really is interesting, and when you have these larger firms involved, they’re all jockeying for position. Just what Meta has done with Muse this time, and now all of a sudden they’re apparently buddies with Apple again, which was not the case three or four years ago. So it’s fun to watch. We’re going to take another break and continue.

[Break]

DENNIS BRAGG: All right, continuing on Talkback this morning as we talk about data centers this hour. We’ll have the Israeli Consul General coming up at the top of the hour. That’ll be interesting. Bob, go ahead. Why don’t you pick up our conversation with Jennifer Warren?

BOB SEIDENSCHWARZ: As I’m listening to what you’re describing here, it strikes me that this whole discussion is still in its early stages, and people are asking: does somebody have a vision for how this is all going to play out? Or is it going to be the wild, wild west, the gold rush? So, Jennifer, help people understand how early this stage of development is, and is there a master plan, or is there going to be a lot of starts and stops to this process?

JENNIFER WARREN: So we really are early, in a way. If you think about when LLM frontier models really hit the consciousness, it was more like 2023. Then in 2024, you started seeing firms like Energy Transfer saying, “These data centers are going to demand more natural gas.” They started seeing this demand growth in natural gas resources, and in pipelines that can help facilitate that, like the one near Abilene, Stargate, and others out west. But the big takeoff, where it shows up in gross domestic product, is in March 2025, and it has continued because we’ve seen all this investment. So we’re in this investment cycle, right?

JENNIFER WARREN: And the things that help constrain the wild, wild west, because it has been a bit of a wild, wild west, and I’m in Texas, and in West Texas it has been the wild west. We have a huge data center queue that’s getting whittled down significantly for all kinds of reasons, including policy. But capital, money, is a governor, a constraint. The market will whittle things down over time. Energy constraints will also keep things in check, because right now, for most of the new-build data centers, it’s the availability of power that’s attracting those sites. It’s driving power-first strategies ahead of everything else.

And that tells you, right there, that there’s going to be a lot of change in energy over time. The grids are going to get more efficient. We’re going to use resources more efficiently. There’s a lot of legacy inefficiency in grids around the country that’s going to get reworked, and we’re going to get these new grids, if you will, and more innovation because of this technology push that sits within these data centers. So I think you do have natural governors, and the market is going to decide ultimately. Even with policy constraints in places, over time, this is how the world is changing.

JENNIFER WARREN: For example, there was an aluminum smelter, and I talked about it at the University of Montana. It had a lot of the jobs and industry in the community. Well, it went away in 2008. Alcoa shut it down, and was remediating the land, and sold it to a developer that bought it in 2021. Now, some of the existing power assets, which used to be coal and were shut down, there are going to be new natural gas-fired plants, which are very efficient and much cleaner than coal and other resources. There’s going to be an advanced manufacturing campus nearby, adjacent to the Samsung chip plant and near the Austin tech ecosystem and cluster.

So you see some of this activity naturally go where these clusters develop, and some of these clusters are established. We have a manufacturing and semiconductor hub in North Texas. It’s evolving. We also have this in Austin. Look at Elon Musk’s Terafab and Tesla in Austin. So you have these things naturally developing.

If you think about it like this: AI in the data center is using high-performance computing power. We’re using the power of computation and digitalization to drive more business, to do things more efficiently. And yes, there will be some reduction to some employment, but those jobs are going to be high-paying as well. So it’s just this evolving economy, is how I see it.

DENNIS BRAGG: I think that’s interesting too, because you mentioned aluminum. The Northwest, especially during the sixties, seventies, and eighties, was one of the leading regions when it came to aluminum production. We had Columbia Falls, we had Kettle Falls. They were all over the place, and it was largely driven by the relatively cheap power we got off the BPA grid, like Grand Coulee and all of the dams. So the market definitely followed where those resources were. Interesting comparison there. All right, we’re going to take a break.

[Break]

DENNIS BRAGG: Okay, we’re back on Talkback, our conversation about data centers, the AI build-out, and the whole kit and caboodle here. Jennifer Warren is on with us, author and consultant, and Bob Seidenschwarz. I’ll get there eventually, Bob. I’m sorry, I’m overthinking it now. That’s the problem. Go ahead. You had our next question here, and then we’ll get to some calls.

BOB SEIDENSCHWARZ: Jennifer, you mentioned the word energy several times. What is, and who is, this energy, in terms of where it comes from and what it’s going to have to require to meet all the parts of the discussion you’re bringing up here on the show?

JENNIFER WARREN: So the energy conversation is a huge thing, but what I can say is that were it not for the shale revolution, we would not be having this AI infrastructure build-out, or AI’s ability to scale and be this great part of GDP. The shale revolution brought all these new resources, oil and gas, around the country, and a lot of infrastructure was built. I can speak to Texas: Energy Transfer and others have really built out the infrastructure that is supporting a lot of the upcoming data center growth.

There will be a lot more natural gas plants built in the future, but they’re actually one of the most efficient ways to produce energy. We do have a lot of renewables scattered around the country, and a high concentration in Texas, but those are dependent on weather. And people say, “Well, we can use batteries,” but batteries have limited lifespans, and these AI campuses need 24/7 firm power. That’s why you’re seeing more natural gas coming online now and in the future. There is a considerable build-out for firm power.

I see shale and the foundation it laid as chapter one. Now we’re in chapter two, and the grids are going to get a makeover over time. I wrote about that about a year ago, and I think this is really true. With the features I have coming out for the November DCEO Energy Edition, you’ll really see how one power producer, Vistra, is thinking about it and doing it, including extending the lives of existing nuclear power plants. There are a lot of efficiencies that can be gained with the resources we have already. Were it not for this AI infrastructure build-out, we would not be getting all this innovation in how energy is produced, moved, transported, everything. So it is kind of creating, you know—

DENNIS BRAGG: Okay, perfect. Let’s get to a couple of our callers. Jeff is on the phone. Jeff, good morning. You’re on Talkback.

JEFF: Good morning, Jennifer. Good morning, Montana. It’s good to be able to chat with one of my favorite two energy-related folks, you and Robert Bryce, who’s been on the show too. You’re obviously well informed, and I really appreciate these opportunities. Here in Montana, it seems like there are really two things to deal with. There are all sorts of things around infrastructure and grid upgrades, but the primary thing for us is, first of all, resource use.

There’s been a lot of hype and/or maybe a lot of legitimate concern, primarily about water for cooling and electricity for powering the plants. I know electricity is one of those consumer-regulated things. But if you could talk to the use of those resources, I think it feeds into my second question, which is the ulterior motives of some of the people, particularly hostile nations, who want to reduce our ability to fully participate in this revolution so that they can leap ahead, and gain on us, and keep us on the defense. So if you could address those two things, basically resource use and—

JENNIFER WARREN: Geopolitics.

JEFF: Any reactions?

DENNIS BRAGG: Go ahead, Jennifer.

JENNIFER WARREN: Yes. Okay. On the efficiency side, there’s a lot of R&D going on in more efficient chips, more efficient networking, energy utilization in the racks, energy between the data center and the grid, everywhere. A lot of these new data centers are closed-loop systems, which use very little water to speak of. In Arizona, for example, there’s a huge Taiwan Semiconductor fab and a lot of data centers. It’s a tier-one data center market, and a lot of the data centers there are air-cooled because of the lack of water. So there’s a lot being worked on in the data center, because they all know they have to have this social license to operate. I’m worried less about that, and in a way I think the public is behind on their understanding of how the innovation is happening in that area of efficiency.

JENNIFER WARREN: It’s all about efficiency, in fact, going forward. That happened with shale, too. Shale has become more efficient over time. You have this period of, “Oh, we’ve got to do this differently. We’ve got to use fewer resources.” So I think that’s what’s happening here too. Absolutely.

Now, on the geopolitics of this, yes, there is this race, no doubt about it. But here’s what’s really interesting: we’re the largest market for AI technology and AI infrastructure build-out. A lot of that has to do with our amazing energy resources, our innovation, our business leaders, our entrepreneurial culture, our markets. We have all these factors that come together that allow us to grow, and we don’t have the heavy hand of government telling everybody what to do, right? Some of these bans and moratoriums are that, but I think there’s also some hedging, just to try to…

[Data centers] create a lot of revenue for the state and local economy, and states and local economies need to figure out how to make this an asset and a benefit to the community, as opposed to making the data center a bogeyman, because I don’t see it like that. So I think we are leading, and I think we need to just keep doing what we’re doing, but also do some of the smart things and pay attention to definite hostile actors.

DENNIS BRAGG: Good point, Jennifer. It seems like every time a new manufacturing operation comes along, local governments have to deal with how this compares and what it replaces. So, real fascinating. All right, we’re going to take another break. We’ll get back to Talkback on KGVO. Let’s go ahead and get to our calls. We’ve got Harry on the phone. Good morning, Harry, you are on Talkback. Go ahead, sir.

HARRY: Yeah, good morning, folks. Jennifer, you talked about the investment and the infrastructure building up, but I just look at what happened back in Eastern Montana. When they talked about the fracking, they had all this money invested. They built up this infrastructure. These little towns popped up. And then when they quit drilling, now they’re producing oil, but those jobs all went away, and so did most of those little towns. So I’m looking at this, unless these data centers will maybe have their own power plant, which would have some people there. But other than that, I don’t see a whole lot of jobs. Just because there’s money being pumped in doesn’t necessarily mean there are going to be long-term jobs. I could see just short-term jobs.

DENNIS BRAGG: All right. What about that, Jennifer? What about the numbers of jobs? I’ve seen that as one of the parts of this debate.

JENNIFER WARREN: How I would answer that is, sure, there are cases where there’s boom-bust, and a lot of it depends on what you’re siting. Are you putting up a power plant? Are you putting up a manufacturing plant? What are you standing up? All I can say is I think over time the market decides that. I can’t speak to little towns that come and go, but there is an example in Childress, Texas, where they developed a solar farm that just sat idle for a long time, building up what’s called congestion costs. And now there’s an AI infrastructure build-out happening there where they’re consuming that solar.

You’re getting at stranded assets, and some of the data center developers are starting to move toward those. In fact, the one in my upcoming feature, in North Dakota, they’ve built some large-scale campuses there, and they’re using some of the stranded wind resources. The ratepayers got a $45 million credit. I think that’s also what’s happening in Louisiana, where some of the underutilized natural gas resources are being used. That’s why there are a lot of data centers being developed in West Texas: same issue, underutilized gas resources that come along with oil production. So I think you’re just seeing this reworking, and there’s going to be a lot of jobs surrounding these campuses being built. Some will stay, and quite a few will stay. It’s just the nature of how work is changing, how the economy is changing.

DENNIS BRAGG: Let’s go ahead and get—

HARRY: I’d just like to—

DENNIS BRAGG: Let’s move on.

HARRY: Okay.

DENNIS BRAGG: Hang on. I want to get Katherine on, because I’m trying to get in as many calls as we can. Katherine, go ahead, and let’s cut right to your question for Jennifer Warren here on Talkback.

KATHERINE: I want to ask about a corollary aspect that you briefly mentioned. You mentioned nuclear power. What about the idea of small thorium reactors, privately developed by these big data center businesses to provide for their power sources? Montana and Idaho have thousands of tons of thorium in the Lemhi Pass area that could be used for this. Is this a possibility to alleviate some of the needs for power?

DENNIS BRAGG: Thanks, Katherine.

JENNIFER WARREN: I can speak to—I know there’s one developer doing these more micro-type, small-scale nuclear modules, and even smaller. They’re called Oklo. They’re not using thorium, but I’m familiar with thorium, and there’s a lot in development with that, for sure. But right now, because of the cost and the requirement for scaled power, that’s why you’re seeing a lot of investment in extending the lives of existing assets. Building a new nuclear plant is very expensive. It’s hugely expensive.

But there will be more nuclear, and definitely there will be some at smaller scales. There’s a firm called Diamondback that is going to be siting some of these smaller Oklo-type reactors, or modules, in the oil patch, because of the need for electricity. So I think you will see some of that. How this evolves is still really up to the market. What can the market sustain? Who’s going to do the R&D investment in these things? That’s how I’d answer that.

DENNIS BRAGG: You do have that research project down at INL, and then of course the one up at Malmstrom, to find out some of those questions. Bob, go ahead. I’ve got time if you want to throw a question in before we go to break.

BOB SEIDENSCHWARZ: Just very quickly. Harry’s point is a common point that people make, which is, what happens when this goes away? Our history has always been uneven, and this is a situation which is probably going to become even more so because of how dynamic it is. Jennifer’s mentioned this several times: the dynamism of all of this is the innovation that takes place. While this may work here, it’s not going to work over there. So this is going to be USA creativity on steroids going forward, and some of it’s going to work, and some of it absolutely will not pencil out as a return on investment.

DENNIS BRAGG: Jennifer, I’m going to have you—let’s go to break. I’ll have you respond.

[Break]

DENNIS BRAGG: And we’re back with Talkback with Jennifer Warren. One more segment here. We’ve got the Israeli Consul General coming up at the top of the hour. That should be interesting. Before we went to break, Jennifer, Bob was also asking about the approach where one size does not necessarily fit all. What may be applicable in Louisiana or Texas, especially because it’s resource-driven, may not work in Montana, or it may be vice versa. What about that diversity of investment, and what we’re seeing develop here?

JENNIFER WARREN: So this idea of moving to power is one issue with the AI infrastructure build-out, but there are other types of businesses happening because of this. I mentioned some of the semiconductor fabs in Texas and Arizona. There’s also a lot of activity in Ohio. There are memory chip plants being built by Micron, and SK Hynix is going to start building some fabs in the United States. So you do see some of this manufacturing reshoring, and from that, it will also spread out and build other businesses. It’s been called ‘building out the backbone of the next industrial revolution,’ and I think that’s kind of what’s happening. People are focused and worried about energy, but I think the market is going to be the governor. Capital’s going to be the governor.

There are a lot of new approaches being developed, even to meet the constraints, I’ve noticed, within this AI infrastructure build-out. There are firms that are evolving to figure out, “Okay, where can I add value, and how can I capture some of that value?” And those are new jobs. These are new jobs and new things that didn’t exist before. One of the things to think about is that whatever happened before, it’s changing, and it may not happen again. There’s going to be new jobs, and there’s going to be a churn of the workforce. In fact, I’m going to a whole workforce event after this, and they’re going to be discussing that. What do we need to do? How do we retool ourselves? How do we take advantage of new opportunities, because they’re emerging? So I think it’s just creating its own energy, if you will.

DENNIS BRAGG: Bob, we’ve got time. It looks like we have one more question. Go ahead.

BOB SEIDENSCHWARZ: I was going to say, this is not your father’s industrial revolution. When you’re talking about reshoring, everything is going to be done in the manner that Jennifer’s describing, using various degrees of artificial intelligence for the very purpose of being competitive, because the rest of the world is not going to sit still either.

DENNIS BRAGG: Now, we’ve got about two minutes. Any parting thoughts, Jennifer?

JENNIFER WARREN: Okay. I’ll just give you this example. This is interesting, and it came out in Meta’s earnings call. Zuckerberg was saying a large reason the Muse rollout happened in the United States is that we have all the infrastructure, if you will. We have the resources. We have the ability to build the AI infrastructure that will house this greater computational power and technology. That’s why they’re not building it out in the rest of the world right now, because there’s not the capacity. Compute capacity is king right now, and that is the data center.

JENNIFER WARREN: So I think you will see a lot of innovation in that space. Part of the innovation is the scale that’s happening, because this IT tech, the guts of the data center, is very expensive. But you’ll see other things in the future too, which gets at data centers of various sizes performing different types of activities, including cloud, for our Netflix streaming and everything. So I think it’s a very dynamic time, and you can’t really say one thing or point to one thing or even shoot one thing in the foot.

DENNIS BRAGG: Well, the thing I’ll be watching, you talk about the reshoring, and I think that’s an interesting conversation for another day, because the cost of equipment is so terrible now with chips and hard drives. If you wanna innovate, if you’re a business or anything that depends on the other end of this compute picture, there’s a lot to cope with, and I’d love to see those prices come down. Hard drives alone are twice or more what I paid just at the beginning of the year. So, very fascinating.

DENNIS BRAGG: Jennifer, we appreciate your conversation, and thank you for your expertise and everything. Real quick, is there anything coming up we should watch? You mentioned an article coming up. Where can we read more?

JENNIFER WARREN: On my site, conceptelemental.com, I will be posting… The print edition comes out this week, and I will have a PDF available on my website. Then the digital versions of these two main features will be coming out around November 3rd and 10th. I also have a Substack called @moreelemental, where I also put work, so if you look there, I will start putting things up. I will have a preview going online probably tomorrow, after this AI workforce conference. That’s where you can find everything. So give me a day or two to get things up.

DENNIS BRAGG: We’ll give you a little time to work with it. Jennifer, thank you so much for joining us. I appreciate it very much.

JENNIFER WARREN: Thanks for having me.

Video references. First video previews trends feature upcoming and second is an overview of the economic impact of the buildout.

Highlight Reel: Jennifer Warren on Data Centers and the AI Build-Out

Read more

Texas’s economy is a place where AI infrastructure and chip investment show up most starkly. In a recent presentation by the Dallas Fed, held in Fort Worth, Texas, new economic data is showing a strong third quarter ahead. Additionally, from bond kings PIMCO, analysis shows that U.S. GDP is actually understating the impact of AI investment.

The AI boom is a host of related economic activities from manufacturing, power infrastructure, software and R&D, and construction and services, to name a few.

Video analysis from the Dallas Fed event focused on Fort Worth, of Dallas-Fort Worth metro area, and some updates on the Texas economy follow. Plus I’ve curated some data and discuss PIMCO charts.

My highlights:

Global and U.S. Growth in the Data, and the Texas Data Center

• True AI impact is understated in GDP because much of the headline CapEx is imported content, which nets out.

• Texas job growth is 1 percent versus 0.5 percent US-wide, and Texas construction jobs are up 3.6 percent versus 1.5 percent nationally, driven by data center building.

• Data center construction was $11.8B last year (just under 6 percent of Texas construction) and is nearing 10 percent in 2026.

• Semiconductor manufacturing is coming to Texas: a new TSM plant with ties to Foxconn, Apple, and NVIDIA; Samsung’s Taylor fab expansion; and a Samsung HQ north of Dallas. The potential TeraFab siting with TSM is still unannounced. They may build in Dallas.

Dallas Fed survey of 267 executives (through Sept 23): selling prices expected up 3.5 percent (from 2.8 in June) and input prices up 4.6 percent (from 3.7).

Inflation drivers: high oil prices, plus services costs (including healthcare) that may account for about half.

U.S. Tailwinds: AI investment, data center construction, oil. Headwinds: labor supply, soft housing, geopolitical and trade uncertainty.

The US is the top AI-equipment import market among developed economies. Taiwan’s exports exceed 30 percent of GDP, and South Korea’s are about 15 percent, driven by memory chip makers.

The equity-market wealth effect and job growth both spread AI’s impact beyond data centers.

PIMCO notes:

PIMCO estimates that the equity wealth channel has provided roughly an additional 0.5–0.7 percentage point of support for growth – roughly 1x to 1.5x the growth impulse of the investment itself.

Their estimates suggest that AI-related forces have provided meaningful support to U.S. growth, even though the direct domestic investment effect is considerably smaller than the headline capex figures might imply.

High import content limits, ie., the chips and other IT equipment components, the direct U.S. growth contribution from AI investment.

However the U.S. simultaneously exports part of the investment impulse to the rest of the world.

The U.S. capex boom is therefore increasingly supporting a much broader manufacturing and trade cycle, particularly across the Asian technology supply chain.

The data center isn’t just represented by AI workloads and spending compute churning out data of various qualities. The connection to commercial real estate, industrial and infrastructural-focused real estate, ie., digital infrastructure, is a missing part of the concept. This is driving some of the financings which are widening the capital stack into new places. We are entering a new day of measurement and disclosure too.

Questions of how much compute (it’s a lot) gets built are relevant. The timelines are relevant and likely extending because of the physicality and supply chain constraints connected to the buildout. Two areas of constraint are power, but memory chips were and are also challenged, plus every input that goes into this.

In new academic research from an NBER working paper, they estimate that a 188 GW scenario of a U.S. buildout completed by 2032 would require investment of $9 trillion, amounting to 3.2% of U.S. GDP. The researchers do a great job looking at the financing aspects of this buildout. I believe this 188 GW scenario is less.

Based on their $41 billion per gigawatt estimate of a new build, a 200 MW facility costs $8.9 billion, more if it is liquid-cooled versus air-cooled. The dynamics of the details matter if you are investing in specific stocks or the AI trade-specific or focused funds, like Spear Alpha (SPRX). The revenue flows are emerging across the landscape. This is a forward-looking snapshot.

The hyperscalers are still largely in the driver’s seat with the buildout however.

The Blackstone digital infrastructure fund (BXDC) has the right idea in the recycling of capital for recently-built hyperscaler data centers. Meta, Google, Microsoft and Amazon are all implying this. Their off-balance sheet financing isn’t a bogeyman, but a notable forking of the IT side and hard assets side of this physical build-out. Investors have choices and different risk styles. These are two mammoth areas as investment needs suggest. Morgan Stanley forecasts “that more than half of the roughly $2.9 trillion required to meet hyperscalers’ incremental compute needs over 2025–2028 will come from outside capital.”^[I] Evidence also suggests that credit investors are repricing the off-balance-sheet financings.

Meta has made choices, reflected by the Hyperion-Blue Owl deal, that indicates seeking to retain more of its asset-light corporate model, not just lower cost financing. It was initially a 2 GW asset with a $30 billion value. This idea was captured in the video article discussing the Blackstone fund as relayed by a former Digital Realty executive that was part of a consortium of financiers for Stargate Abilene in Texas, which included Blue Owl.

The academic research notes that the publicly-traded data center REITs are close to market betas now of 1 versus 0.5. That really should not be the case for Digital Realty, Equinix, and the Blackstone REIT. They are not as exposed to the risk of the entire AI trade, as reflected by their asset bases. It shows you how momentum really lacks nuance, and sometimes knowledge.

The gigawatt (GW) count

While the lion’s share of capex is devoted to the AI infrastructure build out and advancing frontier models, agentic software development is increasing in investor interest. Interest in energy and sustainable technologies are close behind. We will need more efficiencies in those areas as more dense AI workloads require resources, work in which I’m more connected with.

Workloads are expected to increase in the future and require the configuration of chips being developed to evolve differently.

At earnings time ahead, for Amazon, Microsoft, Meta, Google and Oracle, I would expect to see modest growth indications of new spend. Just the inflationary effects alone could level things out. The optics of the election may cause some conservatism even. However, Taiwan Semiconductor reportings are another story, 50% growth year over year.

Here’s why. Scenarios are just that, but for the sake of thought experiment:

From the many operational and announced (planned) data centers according to the NBER paper, I dropped out 2031 years and after as the numbers become less likely.

Then, what is more likely, but still likely exaggerated, are planned projects but with realized capacity as the metric. This accounts for delays, reduced sizes, and abandonments, ie., reality of constraints. (See my video analysis.) This leaves the following scenario, which is still on the higher side of the buildout’s real numbers.

Read more

Data center new builds are expensive.

New builds are costing more, and higher interest rates make new financing pricier. That raises the value of existing capacity and creates room for early, higher-priced capital to recycle into new deals.

Prior work supports thesis.

In summer 2025 interviews for a January 2026 article below, Bill Stein expected early higher-priced capital to recycle and hyperscalers to eventually shed real estate holdings. Stein was also part of the financing consortium, including Blue Owl, behind the first Stargate in Abilene.

I discuss the Blackstone Digital Infrastructure Trust, a REIT, which captures part of the AI trade. The investment ties together hyperscalers, chips, real estate, and energy. It also connects to the MANGOS framework discussed in June.

Reference articles:

Real economy evidence

Several Federal Reserve districts report data-center construction as a leading source of regional activity. This activity extends to manufacturing, industrial real estate, steel, energy, air freight, and computer components. Interestingly, new AI-related infrastructure business models are emerging that include rented third-party infrastructure, hyperscaler GPU rental, and owned infrastructure, increasingly a segment heavily dominated by the hyperscalers and wanna bes.

In a January 2026 publication, neoclouds, largely former bitcoin miners, were discussed from observations across a couple of years of firsthand research. They are spanning these new business models with their competencies gained and evolving the potential of technology they house.

From early September, Federal Reserve Beige Book reporting identifies data centers as a major growth engine for construction, manufacturing, industrial real estate, and power demand in the districts of Cleveland, Ohio; Atlanta, Georgia; Chicago, Illinois; St. Louis, Missouri; Kansas City, Kansas; and Dallas, Texas.

Bottomline: Data centers were a key growth engine as reported by Fed data. In a recent Montana radio show, citizens revealed a different story than headlines portray, reflecting civil discourse and openness to what’s ahead.

Some granularity about the regions:

These regions, noted and discussed in the video, correspond to Federal Reserve Bank districts.

Recent Data on Key Areas of Capital or Investment Spending

From Federal Reserve data, the 2025 totals for the proxies of the AI buildout (minus software) are:

· Data center construction: ~$49.7B for the full year

· Computers/peripherals: ~$270.7B for the full year

Data center construction spending grew steadily (~30% year-over-year), while computers and peripherals spending were growing much faster (~72% year-over-year) to the first quarter 2026.

The AI trade has spread to virtually every corner of the market, though it is more concentrated in places. This is a work in progress, as announcements from various corners continue to roll out.

Increasingly, the use of AI technologies is becoming highly embedded in fields beyond the hyperscalers, though they are top users and builders. In 2026, this investment is increasing with some shifts to the proportionality.

The recent September manufacturing survey conducted by the Dallas Fed is instructive about risks ahead. Inflation was a top concern, followed by geopolitics, domestic policy uncertainty and cost of capital.

Read more Fed…

Read about what people in Montana think.

Below is a breakdown from the video presentation, organized by Fed district.

Read more

One has to look at this buildout from many angles to power through the headlines and noise.

In published work early 2025 and 2026 about the buildout (below), I referenced the concepts about scale. Those ideas matter for this discussion to ground what is happening. Four forces are building atop each other:

• the power of the Internet, with search changing;

• mobile apps and mobility;

• cloud infrastructure growth and changes;

• AI technology bucket, which includes generative AI, machine learning, and many more technologies.

Technology’s Pacing

Importantly, the growth curves of each of the four are creating synergies yet to be determined. Nasdaq reports that the “AI industry” is pacing three times faster than the top three buckets noted above. The expansion of token volume is a proxy for this growth.

In the chart, note that “API-only” is developer use. Most tokens were largely in training output before early 2025, i.e., inference wasn’t on the books just yet.

Whether a token is open- or closed-source, it still requires the same chips and associated infrastructure to produce, and in some cases, more. Models’ competition is fierce. Nasdaq says:

The biggest risk to further upside resides with the closed-model frontier players, their ability to remain at the frontier preserves their pricing power, as more and more open-source-powered AI agents proliferate and require ‘higher IQ’ teachers, coordinators, and supervisors to help them deliver value-add.

Closed models are like Anthropic’s Claude and OpenAI’s GPT-4 and 5. We know various hyperscalers have developed their own too, Google’s Gemini (closed), Meta’s Muse, and others, etc. Amazon uses many via its cloud offerings, including its own proprietary model, Nova.

Agentic AI and tokenomics are a storyline left for another day, but an indication is ahead. According to Goldman Sachs, token use by AI agents is forecast to grow 24x by 2030. Importantly, frontier models determine what tokens should come next. The infrastructure, and specifically memory bandwidth, determines how fast and how cheaply those tokens can actually be produced at scale.

Increasingly, the use of AI technologies is becoming highly embedded in fields beyond the hyperscalers, though they are top users and builders. Even academic research is pushing boundaries and asking questions about the technology’s effectiveness. It’s all new terrain. There will be breakthroughs and failures, but forward movement ultimately.

Select Chip and Digital Infrastructure Dynamics

Some of the “obsolescence of chips story” of fast depreciation is changing. Chips are responding to intellectual property enhancements, such as software upgrades, helping extend their productivity over longer time horizons. Operational innovations are emerging to optimize the mix of chips, networking, power flow, and cooling to squeeze out the maximum possible yields. Data center developers, including those that shifted from bitcoin mining, are innovating in the operational space and identifying new business models. Three distinctive business models have been identified that are proving profitable, but this is a moving target too.

In a recent interview with a developer executive, he confirmed that roughly 75%–80% of the spending on a compute GW is the IT equipment, chips, and networking. The remaining 20% is the land, shell, and related heating and cooling infrastructure. However, as we know, increased memory chip constraints have raised prices, with numbers cited of 15–20%. Top memory chipmakers Samsung and SK hynix raised HBM3E supply prices by nearly 20% for 2026. Even Apple had to raise prices owing to memory chips.

Still, optimization is continuing to increase with each newer generation of chips sourced into the data center market. Below H100s, the blue line is newer chip generations than A100s.

Meaning Through Chips Layer

Further evidence beyond Nvidia comes from rival Broadcom, with a focus on custom silicon. From a recent earnings call, they project ~30 GW aggregate demand across six customers (Google, Anthropic, OpenAI, Meta, Microsoft, and Amazon), translating to a $350B two-year shipment commitment. Interestingly, the dollar content per gigawatt is stable at $20–30B across generations of their offerings. Chips are becoming more powerful, and growth is resulting from more gigawatts deployed, not price expansion. While Broadcom expects to ship $350B in goods, deployment of those chips have their own timing.

Custom silicon, like ASICs and XPUs, are evolving from a niche hyperscaler side project; Broadcom is positioning itself as a main actor in that shift.

A timing mismatch exists in capex spending versus the payoff, with data centers requiring 18–24 months to build and sometimes longer to energize. We know more debt is being issued for the buildout, but that was inevitable.

As Fed Chair Warsh notes:

Will we see a sustained rise in productivity across the economy, and when? Will next-generation models demand more capital, or will models themselves enable a more capital-light path? It’s still unclear where returns on capital will land, and when.

— From Chairman Warsh’s Jackson Hole speech, Aug 28, 2026.

However, as I noted previously, he calls on market participants to inform the debate. This is the role of markets and investors, plus other information intermediaries.

Video clip from 22-minute white paper video walk through and extra economic analysis

Recent Data on Key Areas of Spend

From Federal Reserve data, the 2025 totals for the proxies of the AI buildout (minus software) are:

• Data center construction: ~$49.7B for the full year

• Computers/peripherals: ~$270.7B for the full year

Furthermore, the U.S. imported $93.14B worth of computers alone in Q1 2026 — almost exactly matching the $96.05B total actual investment spend on computers/peripherals calculated from Bureau of Economic Analysis investment data. In other words, nearly all of that quarter's computer investment spending was satisfied by imports rather than domestic production, and only ~$16B was exported back out.

The AI trade has spread to virtually every corner of the market, though it is more concentrated in places. The white paper illustrates where some of the movement is occurring. This is a work in progress, hard to parse, as announcements from the players in the data talk their book. However, the trail is getting warmer.

Jan 2026

March 2025

A Conversation with SMU Cox Marketing Professor Gijs Overgoor, Late 2025. A new version of the research is described at the end.

Original interview, diving into using LLMs to handle data on ad performance

Jennifer Warren: What was the motivation for your study?

Gijs Overgoor: We’ve been studying ad creatives and their effects for a while, and we were trying to figure out the best way to identify what’s actually in these ads. Looking at the existing literature, we noticed that most researchers hire and train a group of research assistants to code TV ads against a set of questions. At the same time, we saw a broader trend of marketing and business researchers using AI for this kind of content coding — mostly with text so far, a bit with images, and now increasingly with video.

That gave us a dual problem to work with. On one hand, we have a huge amount of TV advertising data we’d like to study, but we’re limited by scale — research assistants can do a good job, but they can’t code 5,000, 10,000, or 100,000 videos. On the other hand, AI is improving quickly, especially with multimodal models that can process video, text, and audio, but we weren’t sure how good it would actually be at this specific task. So we set out to see whether AI could help solve that scale problem.

JW: You used Alibaba’s Video LLaMA 2 as your baseline LLM. What did you find — where did it perform well, and where did it fall short?

GO: We deliberately didn’t jump straight to the newest technology. We wanted to see how the tools that are directly available to researchers — ideally open source and easy to use — actually perform. Video LLaMA 2, which combines more traditional machine learning with Alibaba’s open-source Qwen 2.0 model, was one of the top performers in that category.

We ran it and compared its output to our human-coded responses. On the surface, it did a reasonable job — it picks up on a lot of information when you ask it questions. But when you dig in, especially at scale, it misclassifies a lot of information and tends to default to “yes” on binary questions. Ask it “Is this an emotional ad?” — yes. “Is this a funny ad?” — yes. It looks like a reasonable response on the surface, but once you convert those answers into actual yes/no variables, it mismatches broadly, and it’s not easy to predict where it will get things wrong. That inconsistency can have serious downstream consequences if you rely on it — for example, if you’re making ad-spend decisions based on the assumption that the model’s read on an ad is accurate.

JW: You also fine-tuned a model you call “Ad Notator,” which ended up outperforming the human labelers. Can you talk about that finding?

GO: We took the baseline model and asked whether we could train it into a better annotator for ads specifically. We fine-tuned it on the questions and answers our human coders had produced, and it improved substantially — it picked up on the information needed to answer a wide range of questions accurately. Some subjective categories, like emotion or humor, are still somewhat open to interpretation, but in most cases the model agreed with the humans, especially where we had a majority answer across multiple coders.

That’s actually where it got interesting: for ads that had at least three human coders, our fine-tuned model — combined with the human training data — outperformed the very humans it was trained on. Using that broad base of AI knowledge and inserting ad-specific training data filtered out a lot of the noise and the reflexive “yes” answers from the model, but it also filtered out the noise from the humans — the instances where a coder might miss something or find a question ambiguous. That combination really cut down on inter-annotator noise.

JW: You also tested Gemini Flash, which performed better than the off-the-shelf baseline model but still not as well as your fine-tuned model. What did you find there?

GO: It’s worth acknowledging that there’s probably going to be a model that does all of this a little better at some point — that wasn’t really the point of the exercise. The point was to understand what happens when you use a model as-is, before you’ve figured out how reliable it actually is for this task.

To our surprise, Gemini still defaulted to “yes” on a lot of questions. That’s a broader tendency across many LLMs — a kind of sycophancy, where the model is almost trained to please the person asking. Gemini’s answers ended up close to human levels overall, but it still misperformed on a range of individual questions. Our fine-tuned model — despite not being the most current or best-performing open-source model available — was still more accurate than Gemini at moderating that “yes” rate.

JW: So your model wasn’t as much of a “yes man.”

GO: Exactly — and I think we have enough of those already. It’s worth being critical rather than just agreeable.

JW: At a high level, what were some of your findings about what actually makes an ad successful?

GO: The outcome measure we used came from iSpot, a TV advertising platform that lets advertisers pretest and track ads — including things like how often viewers turn the TV off. iSpot maintains a consumer panel of at least 500 U.S. customers who rate ads on a range of items, which rolls up into something close to a persuasion or ad-quality score. They also ask viewers about purchase intent — whether the ad made them more or less likely to buy.

We had our human coders rate the same 2,000 ads on questions grouped roughly into action, information, emotion, and sensory categories — sensory meaning things like whether the ad is visually pleasing or sounds good. A few findings stood out. Action-oriented content — explicitly encouraging someone to buy, to go to the store, to go online — was actually negatively associated with perceived ad quality. Informational content — details about the product, the brand, the price — was positively associated. And sensory qualities had a notably strong positive effect: if an ad was visually pleasing, cute, or had sex appeal, it was well received. Pure emotional or “feel-good” content, on its own, didn’t show the same positive association.

JW: This work could serve as a benchmark for using these tools responsibly going forward. Can you talk about that?

GO: The main takeaway is that you can use AI tools to answer questions about ads, and the output looks reasonable on the surface. But when you actually study how those answers influence downstream measures of consumer response, the results can look very different depending on whether you use our baseline (Video LLaMA), a model like Gemini that defaults heavily to “yes,” or human coders — who, it turns out, are also fairly noisy annotators and can miss or misidentify features that relate to ad response. Our fine-tuned model reduces both that noise and that bias, which gets you more accurate downstream measurements.

So yes, you can use an off-the-shelf model for this kind of task, but you should vet it carefully. You can use our model for similar questions, or — and this is something we’re continuing to work on — future researchers should ideally collect their own data and train a model on it to embed that domain-specific ad knowledge. Even with the most current frontier models — OpenAI’s, Claude, Gemini — it’s still worth running additional tests to check you’re not introducing bias into your advertising research. If you’re using AI to study a phenomenon or to pretest ads, getting this wrong can lead to badly incorrect conclusions — and then spending real money running ads that aren’t actually right for your brand.

JW: Any final thoughts on how this helps advertisers decide where and how to spend their creative dollars?

GO: There’s also a question of whether your best-performing ad — the one people most enjoy watching — is actually the one that drives revenue downstream. In some follow-up work, we’re finding that those two things can diverge. It’s important to ask different kinds of questions to figure out what’s actually working about a given ad for a given audience, and I think our tool can help tease that apart.

For example, in a related paper where we applied this tool, we found that viewers are more likely to turn off the TV when an ad is more informational, even though the action-oriented and informational elements are actually more effective at getting people to the store — it just comes at the cost of how much people say they enjoy the ad. We tend to think about Super Bowl ads as the gold standard because they’re clever and entertaining, but the open question is whether those are the ads that actually drive what advertisers want to drive.

Link to SMU Cox write up and full video interview at bottom. Posted Jan 2026.

Update per Overgoor: “The new version of the paper shifts more toward understanding the impact of advertising, whether large-language models (LLMs) can measure it, and — if we can get at what a smaller sample of humans think — how to correct the estimates obtained from the LLM.

A key issue is that LLMs make inferences about ads that can sometimes have the outcome baked in, which is a problem when you want an independent assessment of the advertisement or the specific element of the ad you’re studying. Humans do this far less. From there, if you have a small human-coded sample alongside a full set of LLM codings, you can compare not just the elements and agreement between the two, but also the downstream relationship each produces once entered into a statistical model.

Our paper then offers guidance on how to infer what the “true” relationship likely is, and what to watch for along the way.”

A few themes covered, based on the timeline’s order, include:

• Cutting through misinformation

• Economic growth and diffusion

• The economics of scale

• Select power and energy

• Jobs and local impact

• Geopolitics and national security

Access to full radio broadcast

Or a commercial-free version:

Key questions and edited responses

Talk Back Missoula radio show on KGVO — September 17, 2026

Q — Bob: Give our audience the foundational overview: what’s actually happening with AI and data centers right now?

A — Jennifer: The white paper report splits the story in two: the technology side (moving three times faster than prior IT waves like the internet or mobile/cloud) and the economic side, which shows up physically in data centers. Data center development is spreading across many regions and is now one of the economy’s key growth drivers — supporting jobs and manufacturing well beyond the data center itself. Energy use and water use are real concerns, but the industry is trending toward more sustainable builds, including a mix of renewables, nuclear and new natural gas capacity for firm power.

Q — N. caller: Pushed back on AI alarmism, arguing AI has existed for decades and fear-mongering is irresponsible.

A — Jennifer: Agreed that AI as a technology dates to the 1950s, at the dawn of the modern computer. There’s a lot of misrepresentation driven by various agendas, and cutting through that noise — separating the technology itself from the data center buildout around it — is exactly what her report tries to do.

Q — R. caller: If AI and AI data centers were limited, how would that affect consumers, and would there be a backlash given how much AI already touches everyday products?

A — Jennifer: Data center growth directly reflects how the economy is shifting — more online shopping such as Amazon, Walmart and more streaming (e.g. Netflix, whose subscription price increases reflect the rising cost of the cloud infrastructure behind it). Restricting data center supply would restrict the digital services people already rely on and are demanding more of. Communities that resist development don’t stop the growth — they just push it to another city, county, or state, along with the jobs and manufacturing activity that follow it.

White paper link

Q — Bob: Can you speak to the “multiplier effect” — the idea that a period of heavy infrastructure buildout eventually spins off broader innovation and opportunity, the way past industrial revolutions did?

A — Jennifer: Construction jobs are temporary and always draw complaints (“boom and bust,” as with the shale boom in Midland, Texas), but this buildout is diffuse — spreading across the whole country rather than concentrated in one place. The jobs that remain carry real economic benefit: data centers pay significant taxes, support an ecosystem of surrounding businesses, and (citing Microsoft’s public figures) generate hundreds of millions to billions of dollars in local economic development. They’re also clean, low-traffic critical infrastructure — not the kind of industrial neighbor people often assume.

Q — H. caller: Wouldn’t distributed power (rooftop solar, batteries) or smaller, distributed data centers built on old industrial sites be easier than huge new plants? Also, how much longer can this level of investment — which he said is “keeping the stock market afloat” — actually continue?

A — Jennifer: Data center scale is a market decision driven by the economics of the underlying hardware: GPUs are extremely expensive (roughly $50 billion per gigawatt by one estimate), so hyperscalers need scale to justify and recoup the investment — hence the push toward large campuses, increasingly sited away from cities. On the “how long can this last” question, she framed the market as being in a “digestion phase,” noting there are real lags between when capex is announced and how it’s deployed.

Q — Bob: How is the Iran conflict — and its effect on inputs like helium — factoring into GPU costs and investment planning?

A — Jennifer: Agreed this kind of geopolitical shock changes the pro forma in real time, which is part of why forecasting this space is difficult. She pointed to Fed Chair Warsh’s comment that the Fed is relying on market participants themselves to help assess how AI-related investment is flowing, since the data isn’t available in real time. She’s been tracking some of this through Fed district data through the end of August, which shows continued spread of growth through the economy.

Q — C. caller: Is China funding opposition to AI/data centers and using bot farms to influence public opinion — a new “space race” for AI dominance?

A — Jennifer: Said this is precisely why she wrote the report — to cut through social-media-driven misinformation. She agreed national security and national competitiveness need to stay top of mind, but argued the right response is to stay clear-headed and keep moving the development forward rather than being paralyzed by misinformation.

Q — J. caller: What about proposals to build data centers in space (e.g., rumored Starlink-related plans)?

A — Jennifer: Personally opposed — space is already cluttered with debris and effectively ungoverned, and she sees off-world data centers as inviting a new kind of race rather than solving anything. She redirected to the terrestrial picture: water use per data center is falling significantly as technology improves, and she expects continued progress toward a cleaner, more reliable grid over time. [New regulations and legislation ahead ill also impact development.]

Q — F. caller: What about offshore/ocean-wave power as a cleaner, water-free alternative to land-based data centers?

A — Jennifer: Called it a complex, separate conversation, but used the question to explain “behind-the-meter” power: hyperscalers increasingly bring their own generation on-site (e.g., natural gas for round-the-clock power, at times paired with solar) because of the interconnection queue bottleneck. There’s no single model — the mix depends on energy source, location, and the specific project, making it “a multivariate equation.”

Q — G. caller: Could small/miniature nuclear reactors (like those used in submarines) power data centers? And are the resulting jobs accessible to local workers, or only to highly specialized talent from places like Silicon Valley?

A — Jennifer: Small modular reactors and micro-reactors are real, funded projects, but still years away from broad deployment. For now, the industry has to work with currently available power sources. On jobs, she pointed to Microsoft’s public commitments to local workforce training, saying that pattern of local hiring and training is showing up increasingly across development locations.

Q — Bob (closing): Where can people go to learn more and do their own homework on this topic?

A — Jennifer: Conceptelemental.com and Substack, “More Than Elemental.” Cited were two features available: The AI Wild Ride (2025) and The New Social Network (January 2026), plus her broader energy-focused work.

The 13-page analysis includes the following sections:

• Technology pacing;

• Financial metrics indicate growth;

• Select chip and digital infrastructure dynamics;

• The factor of time

• Rate of change in the physical world.

A video walk-through of highlights with additional economic evidence of the buildout is included. (Clip below)

Buy the report

The report is intended for individual use and not to be reproduced. Please request fair custom pricing for additional users.

Clip is first minutes of long form video that accompanies white paper.

REGIONAL ECONOMIC INDICATORS ADDENDUM

A clip from the video companion to the white paper

Buy the report

A few select ideas from Main paper:

• The AI industry is currently pacing three times faster than the internet, mobile, and cloud sectors.

• The chip layer is expanding significantly and represents the most expensive part of the AI infrastructure buildout.

• Fed Chairman Warsh noted that market participants must act as information intermediaries; this is owing to technology moving fast and data lags official tracking.

  ---

Below are two important histories: a May 2023 first instance of LLMs/AI and Nvidia’s impact on the Nasdaq, followed by a preview and explainer about the Jan 2026 feature regarding the AI buildout from a Tier 1 data center market, Dallas.

Feature below, but video backgrounder included here.

By request of Seeking Alpha, the impact of $100 oil prices on inflation and the economy are discussed. Occasionally, I’m asked to weigh in on a topic, and this is the burning issue of the moment. I’ll run through some slides based on futures pricing, then look at recent Beige Book data from the Fed, and discuss where I think this plays out going forward.

A short

The full video and summary ideas follow:

Summary ideas:

Futures Pricing

I compare futures prices as of September 9 and one month earlier. WTI is near $97 and Brent is at $101. Dutch natural gas futures are almost $26. We’ve also seen increases across the OPEC basket, the Indian basket, Dubai, and the Murban basket. The Mars basket — a medium sour crude reflecting pricing for key Middle East exporters — is at $115.

A month earlier, WTI was $78 and Brent was $83 — a differential of $19/barrel on WTI and $17/barrel on Brent. Dutch gas was $19 a month ago, up $7 since. This gives a sense of what’s happened over the last month amid renewed fighting in the Iran/Middle East crisis.

Beige Book Data (Reporting Period: Early July–Late August)

Districts generally reported slight to moderate growth. Tourism increased; airlines saw strong demand despite higher fares, showing consumer resilience through summer travel. Auto sales were subdued. Manufacturing picked up, led by defense and data-center-related orders — confirmed by other data. Construction rose on data center projects, which is supporting broader economic growth and jobs.

Eleventh District (Dallas Fed — Texas/Louisiana)

The Eleventh District expanded moderately, outpacing the other ten of twelve reporting districts (two reported no change). Manufacturing input costs were elevated, such as fuel, petrochemicals, fertilizer, copper, memory components. This is consistent with rising CapEx budgets from hyperscalers.

The manufacturing outlook improved slightly despite geopolitical instability. Retail expectations have been positive, heading into the holiday season. Oil price headwinds proved weaker than feared, though sentiment could shift as price increase are still in a “wait and see” environment.

Manufacturing & Oil Patch Detail

Gulf Coast refiners raised output; petrochemical production held stable; fuel exports hit new highs. But E&Ps view current profitability as a transitory geopolitical windfall, not something expected to last; operators are being conservative. Auto sales softened. Supply chains remain highly integrated (Canada, Mexico), so trade tensions matter directly to consumer costs.

In the oil patch specifically: the $100 headline price isn’t necessarily what operators are basing decisions on. Planned production increases reflect a pickup in activity, not pure price-chasing. Upstream shale spending remains disciplined overall — a lesson learned long ago that isn’t changing regardless of politics.

Things to Watch

● Price pressure in fuel feeding through to oil-patch costs.

● Low crude oil inventories tied to the Strait of Hormuz and Red Sea situation.

● Producer concern over lasting damage to energy infrastructure — a drag on efficient oil/gas flow.

● Pipeline workarounds in progress: one route to the Red Sea (bypassing the Strait) estimated complete 2030; a UAE route estimated complete 2027; other bypass routes still TBD.

● 2026 inventory data (Jan–May) shows a clear drag versus historical norms.

Takeaways

Input costs remain elevated and tied to the Middle East conflict, with effects on fuel and materials still working through the system — the full pass-through may not have happened yet. Manufacturing is split, with durables outperforming non-durables (partly AI infrastructure build-out driven). Consumers remain resilient but nervous about prices, and elevated oil creates broader tension.

The US has weathered $100 oil before, but what the “new normal” looks like post-conflict is unclear — any real change will take months to show up. Energy activity remains measured, since producers answer to shareholders and need reasonable returns.

A Wall Street Journal piece noted Tehran’s reduced ability to move goods through the Strait, reflecting real economic stress on Iran. My. take: the Strait’s importance is ultimately secondary to relationships. The Canada-US trade war isn’t helping either, potentially hurting in both prices and jobs outlooks.

Closing Thoughts

Looking ahead: futures pricing a month ago (near $80) didn’t hold given renewed conflict, and volatility is likely to continue before prices moderate. This has no long-term economic upside, and the conflict’s geopolitics weighs on global growth broadly — the US has been more resilient than others, but that won’t last indefinitely, and it’s a dynamic voters are sensing. Inflation is likely to be a key issue heading into the vote, with volatility persisting until then — hopefully giving way to more policy normalcy afterward.

Edited transcript follows with background references.

AI and the Economy

The chairman noted that artificial intelligence — despite being an old concept — is advancing faster than even its biggest supporters and voices predicted just a couple of years ago. That’s consistent with what is being observed: the AI trade just keeps going, but is built on data headed the right way.

He also pointed out that ever-expanding pools of capital are pouring into AI-related infrastructure of all kinds. This ties back to the idea of broadening investment that I’ve written about on Seeking Alpha and elsewhere — big finance has emerged in the game, beyond the hyperscalers. He also referenced a kind of “hyper Moore’s Law,” with scaling laws reshaping both the method and the speed of innovation.

This echoes something I discussed in the context of AI diffusion a couple of years back. Everyone talks about Moore’s Law, but scaling laws are now the more relevant framework — and even those playbooks are being rewritten in real time.

(Interview fall 2025 with semiconductor expert Tyson Tuttle and San Antonio Fed board member about Moore’s law and new developments)

Notably, the chairman mentioned that annualized token sales from the two leading labs, likely OpenAI and Anthropic, have topped $100 billion, a 500% year-over-year increase.

(Add to that, ChatGPT just announced its ads business has surpassed a $1B annualized run rate, less than 200 days since launch.)

Regarding the $100 billion in revenue, Warsh raises real questions: will we see a sustained rise in productivity across the economy, and when? Will next-generation models demand more capital, or will models themselves enable a more capital-light path? It’s still unclear where returns on capital will land, and when.

The Big Unknowns

Much of the analysis I see on [crowd-sourced] Seeking Alpha is grappling with exactly these questions — where the money is flowing, where the capex is going, and who benefits. The chairman also raised the question of who captures the initial benefit: owners of scarce assets like the AI labs, chip makers (which have been on a tear), energy producers (major beneficiaries now and in the future), and cloud providers. There’s also the question of how much value shifts downstream to businesses and consumers — which is already starting to happen.

In the biggest-picture sense, market structure — how value shakes out across all these sectors — is one of the largest unknowns. We’ve heard 2026 capex figures of roughly $725 billion for the top four hyperscalers alone.

[Discussion about chart on capex drawn from an April 2026 study.]

Google and Amazon are taking the lion’s share, with Meta and Microsoft following; Meta is actually outspending Microsoft. Oracle is also in the mix, and interestingly has stayed fairly steady on this measure going back to 2020.

The Fed’s Policy Approach

The chairman described a measured policy approach. Rather than telegraphing exactly what’s coming, he wants to preserve room to step back and form a more solid view before committing. Warsh also mentioned a new task force on productivity and jobs, meant to feed the intellectual groundwork to policymakers going forward.

More broadly, he emphasized that market participants themselves should be tracking real information across the economy. I’ve noticed that this is exactly what analysts, media, think tanks, and universities are already doing. There’s a growing wave of new entrants into this space as information providers and analysts, and that’s only going to continue.

[Chart about compute capacity and market share discussed]

Economic Performance

The chairman noted being impressed with economic performance, which appears to have strengthened. Fourth-quarter growth in equipment and intangibles investment has been the highest since 2021 — when the pandemic was letting up — at 9%. AI has contributed roughly 20% of S&P profit growth over the past year; he attributes more than half of this year’s capex growth to the AI buildup. While hyperscalers account for much of that, it’s showing up in other sectors too.

[I discuss chart on usage from Claude data]

Inflation: The Big Focus

Inflation remains the chairman’s central concern — something I discussed in a video about a month ago. It’s running above target, and commodity prices have been part of that story. This pressure looks likely to persist over the next quarter or two, or three or four — it’s hard to say. That raises the chance of a future rate hike.

Al Value

Specifically, from my research, a Stanford study offers useful context here: the value AI delivers to consumers grew 54%, reaching $172 billion by early 2026, up from $112 billion the year before. That value comes from both cost savings and revenue generation across a few key areas (not exclusively) — customer support (14%), software development (26%), and marketing (50%). Marketing in particular is benefiting on both fronts: new revenue streams and reduced costs, some of it from the same shift from automation to augmentation reflected in the Claude usage data.

The same study noted a key milestone: on July 9, 2025, Nvidia became the first $4 trillion public company. As of the summer of 2026, it has crossed $5 trillion in market capitalization — a striking data point.

Closing Thoughts

Pulling from the chairman’s speech: economic performance is good, but this is new territory with a lot of unknowns. While gains from AI are spreading, we don’t yet know how that plays out over time. Inflation is the one to watch, and a potential rate hike.

Commodity markets bear watching too — as we’ve seen with oil prices amid the Middle East situation, Russia, and the related supply-chain effects. The Iran-Hormuz situation has escalated again today.

Read more

Conceptually, the large-scale investment will become diversified across a pool of capital beyond the Nvidias, Microsofts and Metas. This trend was already starting to emerge through Meta’s financing arrangements in El Paso and Louisiana’s Hyperion for its large-scale projects.

Constructing compute capacity

Last October 2025 at the North America Blockchain Summit (NABS), top public crypto miners discussed the market shifts from crypto mining facilities’ capacity to AI and high-performance computing. At the “Capacity is the New Commodity” talk at NABS, representatives of Iren, Core Scientific, and Riot were present. In an article I highlighted still-relevant takeaways, which I’ll update too:

1) “Interconnection queues were a problem to get sites energized. In Texas’ ERCOT grid, 189 gigawatts were noted as being in the queue.”

Now, that queue is over 400 (much of which is not real). The state is pausing to better review these project. Still, plans are in full swing for large players.

2) “Speed to being energized is more valued now because capacity is in crunch mode. Notably, sites that have power online within, say, 24 months, are much more valuable than those sites that are to be energized in 2030. There’s a premium on existing assets, or capacity, and existing sites that can power up quicker.”

Read more

In a video news round up, topics in the news related to notable energy indicators are reviewed with commentary for context. While geared for an investor audience, the findings are relevant to the U.S. economy at large.

Some Select Summary Points

Data centers as a political flashpoint

● Data center energy and water use has become an easy political target this election season, with two state-level approaches emerging.

● Virginia’s governor is pushing to intervene in the $67B NextEra–Dominion merger (creating $400B in enterprise value, covering ~110 GW of capacity), citing concerns over ratepayer costs, job impacts, state benefits, and that data centers pay their fair share of electricity costs.

● This echoes the Trump administration’s ratepayer protection plan at the federal level.

● Texas Governor Abbott has ordered audits of data center projects given an overstated 474 GW interconnection queue, seeking information on water use, cooling technology, and ownership.

● A batch process meant to streamline the queue — one that drew significant lobbying investment — has been halted, effectively pausing new data center development while ERCOT and the Public Utility Commission review information already in the pipeline.

● Well-financed, established hyperscale developers are unlikely to be materially affected; the pause likely impacts speculative or thinly-financed projects.

Strait of Hormuz / Middle East crisis — market impact

● The Iran/Middle East crisis has not resolved; a brief pause is still at risk of reigniting as Iranian hardliners seek assurances over benefits tied to the Strait.

● US midstream and oil & gas producers have benefited from the disruption; a full quarter of data is now available to assess the export windfall.

● Futures were broadly higher (WTI, Brent, gasoline, heating oil); European natural gas (Dutch TTF) was up 6%, partly reflecting hot summer LNG demand.

ONEOK, an Oklahoma-based midstream firm connects to demand

● Their extensive pipeline network across Texas and the central US, including Oklahoma’s Anadarko SCOOP/STACK play, are well diversified across the oil and gas value chain.

● Q2 2026 (the period the Strait has been obstructed) shows crude oil volumes up ~1.5M bpd, a considerable jump from Q1; aviation fuel volumes up 12%, crude volumes up ~10%.

● In their latest reporting, they noted ~5 Bcf of potential natural gas demand from LNG and data center counterparties, concentrated in Oklahoma/Texas/West Texas, plus LNG and industrial demand along the Louisiana Gulf Coast.

● They have a pipeline under construction connected to Project Kilby, a 2+ GW Chevron-developed data center project in West Texas with Microsoft as the anchor offtaker. I note that these projects take years to move from announcement to execution.

Peer comparisons and funds that have benefited are then highlighted in the video.

Bottom line

● The Strait of Hormuz disruption has been a tailwind for US midstream via higher LNG, refined products (diesel, aviation fuel), and NGL/ethane/LPG exports, as global buyers turn increasingly to the US.

● Expect election-cycle headlines to add noise around energy and data center projects, but the underlying fundamentals should be the focus.

Noted by Finance Professor Feng Zhang of SMU Cox, the impact of leveraged single-stock ETFs and the influence of high housing prices in Seoul were contributing factors that also drove the volatility. Similarly, the U.S. market faces issues like cost of living, inflation, and potential energy disruptions from a Middle East war flaring up again.

South Korea is an advanced Asian economy, alongside Japan. The GDP of South Korea ($1.9 trillion) and Japan ($4.5 trillion) and the U.S. ($33 trillion), set the stage for a discussion on the Kospi.

South Korea Market Meltdown Explained and More

Some other highlights:

Complex ownership structures in South Korea lead to challenges in information dissemination, according to Zhang in a recent interview. He cites ‘work being done by South Korea to improve corporate governance and boost valuations, which will take time.’

The South Korean government’s newly-announced $950 billion AI initiatives will be spent over time.

Execution and delivery on promises from tech hyperscalers, despite the market’s reaction to announcements, are what matters. Investors should focus on the long-term plans and capex execution of companies, rather than short-term market fluctuations.

Share

Subscribe now

Bonus clip of SMU Cox Finance Professor Feng Zhang on Korean stock market, part of a longer video interview dated 7/27 on his research about global stock market returns.

Additionally, some overall macroeconomic risks are noted regarding the continued military campaigns between the U.S. and Iran and its implications for energy.

My overall take on U.S. stock market earnings: good news is to generally be rewarded, though not always — some results have “backfired” when the market drew the wrong inferences. Continued transparency on strategy with execution will keep paying off, as it has for Apple.

The top four hyperscalers (Amazon, Microsoft, Alphabet, Meta) are spending $725 billion this year. Google and Amazon have both signaled capex increases for 2027.

Apple and chipmaker Nvidia are now neck-and-neck for the world’s largest market cap, both just under $5 trillion. As of 7/20, behind them: Google ($4.2T), Microsoft ($2.9T), Amazon ($2.6T).

Apple’s re-rating is likely owing to its integrated hardware/software ecosystem and its announced of a chipmaker acquisition — the same “integrated” logic hyperscalers get credit for, applied to Apple.

Also noteworthy is hyperscaler capital spending.

• The top four hyperscalers (Amazon, Microsoft, Alphabet, Meta) are spending $725 billion this year.

• Google and Amazon have both signaled CapEx increases for 2027.

A new partnership was announced between Microsoft and AMD, indicating how collaborative and fluid this ecosystem is.

• Microsoft announced a partnership with AMD for new “Helios” racks — a full-stack, integrated system running on it Azure platform.

• In December of 2025 at an event in San Francisco, chief executive of AMD Lisa Su commented that ‘AI is not a bubble.’ The announcement from yesterday was a bump for AMD’s stock and a sign of broadening competition among chipmakers.

In the video, I discuss some of the holistic factors supporting and constraining the buildout. The erosion from inflation is more of a problem, and it can manifest through a number of channels. Worth noting: the United States is largely a consumer economy, so the consumer matters. However, new investment on a different scale is undeniably impacting GDP and growth.

Update: In a new announcement, Apple says it will identify a chip manufacturer to acquire. This synchs up to the landscape portrayed in the video.

Key ideas mapped to timestamps:

Macro & Inflation

• 0:02–1:08 — Inflation is a bigger concern than AI CapEx “blowing up”; U.S. is a consumer economy, so consumer health drives the whole cycle.

Chips & Memory

• 1:10–1:53 — Memory chip costs are driving up the chips layer of the five-layer stack, adding a “speed bump” to build-out.

• 1:54–2:43 — Chip/memory supply chains are shifting globally

Iran War / Energy

• 1:54–3:48 — Iran war raising energy costs (gasoline, diesel), felt most in Asia; this inefficiency from supply chain reflows the market, which is yet to be understood.

• 16:08–16:43 — Strait of Hormuz strain shows oil market resilience but is a shared global chokepoint that can feed inflation.

  ---

Hyperscaler Capital Spend

• 3:17–3:48 — Hyperscaler debt issuance hasn’t strained credit markets

• 5:28–6:10 — ~$725B slated for 2026 across Google, Amazon, Microsoft, Meta, with nuance described.

• 6:49–7:22 — Microsoft: $190B spend; $25B of it from memory costs; CEO signaling a slowdown.

• 7:22–8:38 — Amazon: $200B spend (mostly AWS-directed); OpenAI and Anthropic revenue expected, incl. 5GW of capacity.

• 9:11–9:47 — Meta notes memory costs + capacity tightness as constraints

• 13:54–14:29 — Google: integrated across the full stack; cloud revenue +63% YoY to March 2026.

Market Structure / Stack

• 9:48–10:24 — Five-layer AI stack (Jensen Huang’s cake); application layer showing more stress of late.

• 12:07–13:19 — Hyperscaler + leased data center share: <50% (2023) → 57% (2025) → 74–75% (2029–30 projected).

Closing

• 15:05–15:36 — Constraints by layer

• 16:43–17:21 — Build-out is a secular 5–10 year force

More analysis and research:

Read more

Why did they come?

European explorers traveled across the Atlantic Ocean to North America attracted by the promise of gold, silver, and other valuable commodities. Beginning in the sixteenth century, short-term voyages gradually evolved into more sustained attempts at colonization. As Spanish, French, Dutch, and English settlers competed with one another for territory long occupied by Indigenous communities, disputes arose, often erupting into armed battle.

Missionaries and colonists from France and Spain came seeking converts, with Protestants from England interested in the same in the Northeast. Wealth and power became concentrated in elites who owned large plantations. Fortunes were made through investment in the profitable “triangle trade,” in which raw materials from the colonies (such as sugar, rice, and lumber) were exchanged for manufactured goods from Europe and New England (such as rum, guns, and cloth). Objections to the slavery were rarely made until the mid-eighteenth century.

It begins

In the 1760s, increased taxation and government interference fueled resentment in Britain’s American colonies. Hostilities smoldered for a decade before erupting into armed conflict in 1775. Many colonists wished to avoid war, but representatives of the thirteen British colonies who formed the Second Continental Congress ultimately decided that it was inevitable. On July 4, 1776, they declared independence from Britain while also attacking the time-honored system of privilege conferred at birth on an elite class. Their assertion that all men are born equal and endowed with inalienable rights has served ever since as an inspiration and a challenge to countries around the world, including the United States.

Indigenous nations were not part of the 1783 peace negotiations.

Creating Culture

After gaining independence, the United States began the work of developing an independent culture. Writers, artists, performers, and intellectuals emulated the traditions they had inherited from the Old World. Nationalism sparked a challenge of deriving original subject matter from the history, land, and inhabitants of the United States.

Thought leaders, and leaders otherwise, drew inspiration from the natural scenery of North America and the vast continent’s dramatic differences in topography, climate, and wildlife. The landscape was romanticized as an untamed wilderness, which ignored thousands of years of cultivation by Indigenous communities. Regional differences in the manners, customs, and language of the Northeast, South, and West infused many of these works, as did democratic values, such as individualism, optimism, and vitality. These creative and intellectual achievements helped to define the emerging features of the United States.

Onset of industry and commerce, and unrest

In the decades after 1820, the United States experienced a period of rapid change. This was the era of canals and steamboats, the beginning of the railroads, and the acceleration of commerce and industry. Enthusiasm for improvement led to investment in education and religiosity.

The ideals of democracy and egalitarianism became a leveling force. In 1828, Andrew Jackson ran for president as a political outsider from the West. He had served as a first member of the U.S. House of Representatives from Tennessee. He sought to wrest power from Northern elites.

In 1830, legislation approved caused Indian tribes west of the Mississippi River on a grueling “Trail of Tears.” The southern lands they left behind became vast plantations supplying cotton to textile mills in the North. Marginalized groups such free Black men and women sought change through religious and secular organizations increasingly by 1860.

Territorial Identity

For much of the nineteenth century, the United States pursued the goal of exploring, settling, and governing the vast country that lay to the west of the original coastal states. Many commentators believed that the United States was preordained to expand westward to the sea and become a continental nation.

Faith in this “manifest destiny” (a phrase coined in 1845) prompted aggressive expansionist policies that often led to violence. Through a bloody war with Mexico over Texan independence (1846–48), the United States gained immense tracts of land stretching to the Pacific Ocean. These developments came with consequences. The disputed status of slavery in the newly acquired territories brought the nation to the brink of civil war. Legislative compromise postponed the crisis—but only temporarily.

Debates about the shortcomings of the United States as a just society gained momentum during the mid-nineteenth century. Some campaigned for or against the key political and social issues of the day, including temperance reform, abolition, and women’s rights. These reform movements coalesced around the institution of slavery and states’ rights. Opinion in the industrialized North increasingly diverged from Southern leaders. By the mid-1850s, the opposing camps had become too inflamed for agreement, and episodes of armed conflict ensued.

This brings us to the Civil War of 1861-1865, which President Abraham Lincoln steered the country through, only to be assassinated shortly after being elected a second term.

Special offer for the 250th…until July 11, 2026

Get 25% off for 1 year

The Frontier of Texas Begins

For my part, this broader part of the storytelling is finished, and I will set the general stage about Texas, where my family lines converged in the beginnings of Texas. The remaining curated narrative is from the National Portrait Gallery, and it has a purpose.

American revolutionaries of Virginia came to Texas from Tennessee in 1821 and 1829, respectively. Sam Houston (Rockbridge County, VA, 1793-1863) was Texas’ first president as the Republic of Texas. He played an outsized role in early Texas history, fighting in the 1835 Texas war. The Republic was annexed to the U.S. in 1845.

Other Texas notable, Stephen F. Austin of Wythe County, VA (1793-1836) was part of the first settlement to Texas from his father’s land grant. He kept Texas under Mexico’s rule, but joined the Texas war for independence in 1836. He is considered the founder of Texas. Also, renowned explorer Davy Crockett, a Tennessee congressman (1827-31, and 1833-35) joined settlers who were promised land in exchange for fighting for Texas independence. He died at the Alamo in San Antonio, Texas in 1836.

The rest is history for another day. In the meantime, be free.