Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
What the AI datacenter build out looks like from the ground up image

What the AI datacenter build out looks like from the ground up

S4 E42 · Bare Knuckles and Brass Tacks
Avatar
1 Playsin 8 hours

What happens when a community votes no…but the #AI datacenter construction starts anyway?

That is not a hypothetical. It’s what happened in Saline Township, Michigan, when a $16 billion OpenAI-Oracle data center was rejected by the local planning commission, rejected again by the township board, and broke ground weeks later anyway. The developer sued. The town settled. They had no real choice.

Sharon Goldman has been covering the AI data center buildout for Fortune — not from boardrooms, but from township halls, planning commission meetings, and rural communities that had never imagined something like this landing in their midst. What she’s found is a story that the technology press largely isn't telling: the buildout is a bottom-up crisis dressed up as a top-down triumph.

The numbers tell part of it. Saline Township received $14 million in community benefits from a $16 billion project, against an annual budget of $1 million. In Richland Parish, Louisiana, the land where Meta's Hyperion facility now sits was once pitched for an auto plant that would have created two to three thousand permanent jobs. The data center is promising 500. The construction workers are mostly from out of state.

And the justifying ideologies — the race with China, the national security imperative — has no finish line. This race has a vague one-upsmanship and a $700 billion spend with no clear end in sight.

What Sharon sees coming, and what she thinks the press is missing, is the backlash that is quietly becoming a political force — showing up in recall elections, in governor's races, and in the kind of conspiratorial thinking that emerges when people have lost trust and no longer believe that democracy is working for them.

You can read more of Sharon's reporting here:

Recommended
Transcript

Impact of AI Data Centers on Communities

00:00:00
Speaker
You know, in other ah industrial projects, the the community might feel like it benefits them more in terms of jobs, for example. An auto plant might have thousands of jobs.
00:00:12
Speaker
Here, the construction jobs might be a lot to start, but those aren't always local. And again, that doesn't necessarily translate into a lot of jobs down the line. There might only be a few hundred permanent jobs.
00:00:28
Speaker
And as you say, ai is a little bit amorphous. Like people don't even realize that all of the data centers up until now really allow us to use email and social media and Google Docs and Microsoft Word and all the things we use.
00:00:45
Speaker
So when it comes to AI, it feels even more amorphous. Like maybe they use ChatGPT, maybe they don't. They certainly may not feel like This is the future where we have to bring these data centers to my community.

Introduction to Podcast and Guest

00:01:07
Speaker
Yo, this is Bare Knuckles and Brass Tacks, the tech podcast about humans. I'm George K. And I'm George A. And speaking of humans, today we are talking with Sharon Goldman, an AI reporter with Fortune magazine. And this episode gets really on the ground among humans in rural communities because Sharon has been working on a long running series for Fortune. about AI data center build outs. She spent a lot of time in city council meetings, planning commission meetings, and hearing from people locally about what it's like to have these projects, these $16 billion dollars projects, these mega data centers bigger than Central Park in Manhattan just kind of show up in their communities.

Sharon Goldman's Research on AI and Rural Areas

00:01:49
Speaker
And yeah, we really drill down into the human impacts there.
00:01:53
Speaker
Yeah, i mean, this is a real professional investigative journalist doing, i mean, call it being religious agnostic, but still the old saying, doing God's work. yeah She really is doing God's work around this and just really insightful person. i you know, I see a lot of the news. I see a lot of the hype, but having someone who's actually there in the room, um I found this episode is as insightful and educational as hopefully the audience will as well.
00:02:18
Speaker
Yeah. So let's turn it over to Sharon Goldman.
00:02:25
Speaker
Sharon Goldman, welcome to the show. Thank you so much for having me, George K. and George A. Hello. We're very excited for this. um You've done some excellent reporting on the data center build out, which has been ah talked about a lot on this show, and we'll get into various aspects of that. But um let's start kind of on the ground. I think a lot of the press typically starts from the announcement, which is kind of between corporate leaders, but your reporting has really been in township halls, city council meetings, planning commission meetings, ah not boardrooms, right? So can you tell us a little bit about what is the data center story, quote unquote, look like from that bottom up view?
00:03:12
Speaker
The data center story from that bottom-up view is really around how data centers have legitimately been around for 15 years, ah longer even, you know, through the cloud computing boom. um The AI boom is now a much larger scale of a data center. We're talking data centers that are that are much larger in size and and need lots more land. They also need a lot more energy, a lot more electricity and ah in some cases, water. um And so builders, developers, and AI companies are targeting more rural areas because that's where you're going to find a those big tracts of land. You're also going to find them next to high-voltage transmission lines with access to water. So the story of AI data centers is not the old-school data centers which you might find in a city. And perhaps... You know, they've got they got bigger over time, but maybe the size of a Home Depot, let's say. But here, these are the size of, you know, dozens of football fields, if not much larger. You know, Meta's Hyperion is bigger than Central Park in Manhattan, to give a sense of scale. Exactly. So it's a far larger scale, typically targeting rural areas, areas where the residents may not ever have imagined that something like this would come to their area. And you're talking about very small towns, which have typically part-time local officials who are very inexperienced in dealing with a project of this size. And that creates all sorts of issues in the local government, um among residents.
00:04:55
Speaker
And just kind of a shock, I would say. Yeah, I remember when the first major one, I remember it was at the NSA about their behemoth of a fortress in Utah.
00:05:07
Speaker
I remember when that happened like maybe 20 years ago. And I saw that thing and I was like, this is... massively futuristic. Um, I, I didn't, I should have known have been the start of a trend, but I didn't think that it would go, you know, taking over was all of Vermont or or West Virginia, George, which state is it?
00:05:28
Speaker
Uh, yeah. I mean, basically all of Northern Virginia is like Amazon East one, like Ashburn, Virginia is just like data center town. Yeah. So, so what's always on in Salien Township specifically, the community, as you mentioned, the, the, the counselors, part-time counselors are usually like, you know, the mayor is a farmer and he's a full-time farmer and happens to be the mayor. And that's the thing. And that's very common in Canada. It's the same thing, but the community voted no, but construction started anyway.
00:05:57
Speaker
So walk us through that happen because is there any real power and governance here?

Challenges Faced by Rural Communities

00:06:03
Speaker
I think what I really got out of that particular story was that zoning laws really differ between states um and even local areas. So in Saline Township, Michigan has a particular zoning law that is called exclusionary zoning. So you can't, you know, if you don't have, say, um
00:06:28
Speaker
Like this was an agricultural township. They didn't actually even have zoning, any zoning in the township for industrial. So there wasn't anywhere in the township where an industrial project could go.
00:06:42
Speaker
So in this case, when this data center project came in, the developer sort of took advantage of that and the landowners by saying, by rejecting ah our proposal for this data center, you are...
00:06:56
Speaker
you know, this exclusionary zoning comes in. You're saying that no industrial can come in because you don't even have an industrial zoning law. um So they were able to take advantage that to sue the township.
00:07:09
Speaker
And then the township, which is very small, you know, wouldn't be able to afford to fight that in court. um In addition to the fact that the developer had said that they could go a different route by aligning, say, with the University of Michigan, which doesn't have to abide by the same zoning laws, they agreed to settle.
00:07:32
Speaker
Yeah. I've got a lot to say there. It's just like, it's just like, uh, we're going to march in with this sort of legal cudgel. Right. And just, we have all the time in the world and all the resources versus like a, they, they, they showed up already ready for now, basically. Yeah.
00:07:51
Speaker
yes and They did show up already for now. It seemed like they had all their, you know, documents in hand ready to go forward with the lawsuit. But what I did find interesting is that people also blamed the the board.
00:08:07
Speaker
But the board pointed out that, again, ah you know, they felt that the the developer would be able to go forward anyway. And in a way, settling allowed them to keep some control over the project and put forth a series of rules and and for the developer to abide by um So in a way, it was it was the one way that the local area could actually take some control over the project. But it really just shows how you know local and really drilling down to the to the most local community and and government that that these projects get to. You're you're not talking about a large city government. You're not talking about state government. or Although in this case, and in many cases, the state government is involved in promoting these data centers. They want the data center to come in. In this case with Saline Township, Governor Whitmer of Michigan,
00:09:05
Speaker
wanted OpenAI i to come in. um And Saline Township just, again, happened to have that perfect site, the land that's flat and contiguous, that has the high voltage transmission lines, access to a highway, access to electricity, access to water. Yeah, I think it's really worth noting this rural-urban divide,
00:09:33
Speaker
I think, you know, in a society that's already been polarized through algorithmic media, you know, this these fissures sort of get a little bit deeper, right? Because it it strikes me that you're asking communities that are probably not seeing the quote unquote benefits of AI, which we'll talk about a little bit later.
00:09:55
Speaker
But like if you are a farmer, are you using like chat GPT on a daily basis? No. So if you can't see this trade off, it does feel like this force just coming into your town to plant something in the middle of it that is for a technology you may never use.
00:10:14
Speaker
You know, I get why that feels. um Like a false choice. And in your reporting, you mentioned that Celine did get $14 million dollars in what's called community benefits.
00:10:27
Speaker
But in terms of scale, it's a $16 billion dollars project. $14 million is great when the annual budget of the town is $1 million. But i guess I guess is like, is that a benefit or like a consolation? Because if we do the math, that's like, I don't know if it's a one-time benefit, if it's annual, but it's still like 0.09% of the project cost.
00:10:50
Speaker
You know, it's like,

Financial Benefits vs. Project Scale

00:10:52
Speaker
what does that mean? Right. I mean, to be fair, that was part of the settlement and kind of the upfront, you know, supposedly over the years, they would also be getting benefits and taxes and things like that. But that's that's far down the line. I think for a lot of these communities, I found this in every place that I've been in Richland Parish, the Meta site and in Louisiana. i have a site that I'm finishing up a story now outside of Waco in Texas.
00:11:20
Speaker
ah in Arizona, a site that has yet to be built. In all of these cases, it's also the speed at which these have come to town. um You know, from not hearing about it, literally never having heard of a data center because in previous cases, waves of technological advancements, these ah data centers at a smaller scale have typically been in areas like Virginia, you know, in in Data Hall Alley or whatever they call it. um And suddenly they're there. Like in in Louisiana, a lot of people mentioned it to me. It felt like yesterday it wasn't there and then suddenly it was um because they typically don't hear about it until it's
00:12:01
Speaker
you know, the the some of the deals and arrangements have to to push it forward have already been finalized. and And so just the pace of how it moves through tends to feel to residents very fast and that they don't really have time to wrap their minds around it. And again,
00:12:20
Speaker
You know, in other ah industrial projects, the the community might feel like it benefits them more in terms of jobs, for example. An auto plant might have thousands of jobs.
00:12:32
Speaker
Here, the construction jobs might be a lot to start, but those aren't always local. And again, that doesn't necessarily translate into a lot of jobs down the line. There might only be a few hundred permanent jobs.
00:12:48
Speaker
And as you say, ai is a little bit amorphous. Like people don't even realize that all of the data centers up until now really allow us to use email and social media and Google Docs and Microsoft Word and all the things we use.
00:13:05
Speaker
So when it comes to AI, it feels even more amorphous. Like maybe they use ChatGPT, maybe they don't. They certainly may not feel like This is the future where we have to bring these data centers to my community.
00:13:19
Speaker
i think I think it kind of stems into two there's two separate kind of points to go with this. The first is that I think um nationally, America is trying to compete directly with China. And I think, you know, America's advantage and in in data and in AI is on creativity, is on algorithmic creativity, right? Like the research case capabilities.
00:13:40
Speaker
um But they're trying to compete with them as well on muscle, on sheer compute. And I think trying to compete with a communist country that can nationalize every resource with complete authoritarian control versus a democratic country where people's opinions and the able to standard of living matters, um that's probably a big crux of the problem. I'm just listening to you talk. That's the first thing that comes to mind. The second is that you know the jobs argument is doing a lot of political work here. And I know that's what they love pushing when they get into these towns.
00:14:08
Speaker
But, you know, when the construction jobs end, if they even come, since half these data center investments seem to end up being canceled or delayed, how does the permanent employment picture at one of these facilities actually compare to the kind of plant that used to anchor a town, like an auto plant or a mill or a manufacturer?
00:14:28
Speaker
Yeah, that's what I mean. So just to give an example, in Richland Parish in northeast Louisiana, that's the Hyperion site for Meta, That land that it sits on was originally pitched for many years to bring an auto plant to town. That would have brought a couple thousand jobs, I think, two to three thousand jobs, permanent jobs. That never transpired. They never were able to get an auto plant.
00:14:54
Speaker
This ah facility, even with its many expansions, I mean, the the first wave that they're talking about, they've talked about maybe 500 permanent jobs.
00:15:05
Speaker
There are about 5,000 to 8,000 construction jobs there right now. There are lots of construction workers. Not all of those are local, though. They're bringing a lot of people from outside. yeah As far as long-term permanent jobs, it does not...
00:15:23
Speaker
ah carry that that you're going to get that kind of community jobs that you would have got with an auto plant. However, the developers pitch that there's a lot of ancillary jobs. Once you get the big data center in town, other industry comes to town. That's already happening in the Richland area.
00:15:43
Speaker
There are other big projects supposedly coming to town, including other data centers, by the way. um and then you have in my in my article about the Richland Parish site, there are ah there are legitimately lots of other ancillary business that are growing to serve the construction workers.
00:16:01
Speaker
But again, because there aren't really... any or very few of these mega AI data centers that are completed, you can't really tell what's going to happen after the workers ah construction workers leave.
00:16:18
Speaker
For example, I think the one that is farthest along is Elon Musk's ex-AI site in Memphis. And, you know, that remains to be seen because they're still building a second site. and I think it it hasn't all played out even there.
00:16:34
Speaker
Yeah, and I think to George's point about the way those plants used to anchor town, whether it was the steel plants in Western PA or the auto plants in Detroit, and even auto plants today in ah Memphis and in Mississippi,
00:16:49
Speaker
it the the promise is that that full-time job gives families, you know, the income to then, you know, send the next generation to college. Like we had a whole, we have whole generations of people who are like, oh, I'm the first and they came from working class towns. But if the jobs are just this sort of temporary injection into the local economy, maybe it's hospitality. Sometimes they have to put the construction workers up in a hotel for and a year or whatever.
00:17:17
Speaker
And then it just sort of like pops that is definitely not as sustainable.
00:17:39
Speaker
And I just brought up this earlier about these rural communities. They're being asked to carry the load for a technology that they may or may not be participating in.
00:17:52
Speaker
i want to pull on a thread here in terms of
00:17:59
Speaker
This isn't quite in your reporting, but maybe you've sort of heard a little bit on the ground about what does that bargain feel like? Because we work in the tech sector where there is a lot of anxiety and hand-wringing about job replacement, right? So...
00:18:15
Speaker
mostly urban centers, suburban centers. We don't see the data centers. We don't hear the construction noise. We don't hear in the case of some of these places that have gas turbines running, you know, 24-7 to power the data center.
00:18:31
Speaker
i kind of want to think a little bit more about that divide because I also think we don't have as many ingestion points in the U.S. economy. For the most part, the U.S. AI sector is sort of racing towards this vaguely defined AGI ASI goal, but no one really knows what it means. And if you're not seeing the material benefits in your life, it's sort of like, what why? It always feels like these things are coming at us from like the same 10 people, essentially.
00:19:02
Speaker
I guess I want to get your sense of that sentiment on the ground. I know that they feel like it's come out of nowhere, but is... ah Is there any sort of reaction to this idea of we have to beat China? This is the most transformational technology since fire, like all these, the hyperbolic claims that are being used to justify $700 billion dollars of data center investment across more or less the magnificence of it.
00:19:31
Speaker
Well, this actually leads directly to something I've noticed recently, which I'm planning on reporting on but haven't yet. And that is I really do see that in this vacuum of trust, I would say, where ah these AI data centers are coming to communities so quickly.

Conspiracy Theories and Transparency Issues

00:19:49
Speaker
Right.
00:19:49
Speaker
at such a large scale, without as many jobs as people might assume, and with such disruption and chaos um expected, even if the project hasn't arrived yet, I'm seeing a tremendous amount of conspiracy theories. I'm seeing online in these Facebook groups a tremendous amount of suspicion and distress talking about what are these data centers? What kind of data are they holding? This is surveillance centers. This is for mass surveillance. This is for population control.
00:20:24
Speaker
This is This is coming to farms so that farmers can't grow food anymore and will be under the thumb of the government. You know, the kind of, you know, it's it's getting a little crazy out there, which is a shame, I think, because there are some legitimate concerns that residents have that are real. But it's also, I think, fanning the flames of conspiracy theories that are true.
00:20:48
Speaker
real either. Which I mean, I think you point out vacuum of trust where there's that gap. Yes. Something rushes in to fill it. I feel like there's just not a lot of transparency either. That's part of it. um A lot of these projects have historically been negotiated behind closed doors with non-disclosure agreements. They don't get announced until it's you know, closer to reality. And therefore it comes as a surprise to the community, but it's not a surprise to you know, state officials let's say, or the developers and and tenants themselves.
00:21:27
Speaker
I was i going to say, it sorry, in my mind, i look at this in recent meetings, I've seen like the comparisons between like, they're calling this now the second Gilded Age because of the oligarchs, the tech oligarchs running all this, and then the government's really falling in line.
00:21:41
Speaker
the first skill of the age, you had Carnegie, you had Rockefeller, you had Ford, you had these big industrialists that, yes, made obscene amounts of money, but it also led to an industrial revolution that directly benefited the majority of working class society, right? It put people back in the cities. It created a lot more technological advancement. It generally...
00:22:01
Speaker
you know, without, there's issues, you can go back to this as a separate podcast, but it generally made life better. And so people saw the direct benefit of this. So these people were put up on a pedestal little bit. In this case, in this gilded age, if fighting China is the only thing that really is to gain, quote unquote,
00:22:21
Speaker
um I don't think people are seeing the benefit because what they're seeing is cost of living is going up, the environment's being destroyed, their health is going down, the ability to to understand, to to create you know any kind of equity for yourself, whether it's a household, buying a house is next to impossible for folks our age now. If it's trying to save for retirement, pension funds are almost worthless if they're even available to people anymore.
00:22:46
Speaker
I think that's a big part of the problem that you're speaking to. And I think that's where government's failing to kind of serve the people because again, we could say because of what's that, Citizens United, everything can be bought now.
00:22:58
Speaker
So there's no real advocacy for the people, which is I think where the problem lays. But to digress, I just, you brought up some really good points and it triggered like, Hey, there's a much bigger conversation at play that needs to be considered.
00:23:11
Speaker
If we look at no sustainability, right, and we talk about benefits for local populations, the hardware inside these facilities seems to depreciate very fast. I think your colleague Sean Tully wrote about AI hardware that hardware that's essentially worthless in about three years.
00:23:26
Speaker
So what happens to a community that's built expectations around this behemoth of a facility when that cycle turns? I don't think that um that's quite right to say that they depreciated three years and are worthless. I think that that it's not quite right because even today, NVIDIA chips that are, you know, three to five years old are still being used for many workloads. I think it is true that there will be constant upgrading to these sites um as far as...
00:24:00
Speaker
the chips but also the networking and and all the other infrastructure that's around it. um and that And that's part of the issue. it's kind of it It will kind of become like this ongoing project. And in a way um developers sort of point to that as a good thing because then you'll always need some construction workers around. um But of course, again, it's sort of...
00:24:28
Speaker
you know is on I feel like a lot of this is around uncertainty. there's just there hasn't It hasn't been around long enough to play out where people can see really what happens with these data centers. Even with old school data centers, the smaller ones that are in more urban areas or data center, you know, areas that are, have lots of them, like in Virginia.
00:24:52
Speaker
um I've been told that it was only really five years ago that communities started complaining and criticizing and rising up about it because there were just so many of them. It just...
00:25:03
Speaker
was overloaded. So here too, I think, you know, now that we're a year and a half into this build, that those criticisms are rising, but it isn't clear what's going to happen as far as sustainability, as far as the environment, as far as electricity and water. So that uncertainty feeds the mistrust as well.
00:25:27
Speaker
Yeah, that's a, that is a good point. um Because it's a lot of vague promises, right? Again, like most transformative technology since fire or whatever. ah George and I have have been pretty clear in our preference for smaller focused models that, you know, apply to a particular use case. We had a a founder earlier in the year on the show using very small data sets and shared compute in a European ah cluster to help small farmers like identify bad crops and stuff like

Preference for Smaller AI Models

00:26:02
Speaker
this. It doesn't require um a data center that's bigger than Central Park. And I think the more we can think about ways that you could
00:26:12
Speaker
make AI feel more real in people's lives than and and much more real than machine learning recommendation algorithms on Netflix, for example, but something like tangible is going to is going to be helpful. So we have mentioned China, I don't know, three times already.
00:26:27
Speaker
i think the race for AI supremacy is super vague, even though it's doing a lot of the work here. um You know, the space race gave us a lot of technology, gave us the internet. um We've also had booms in technology like Sunbeam, the home appliance company, came out of World War II, Teflon, ah nylon, all sorts of things.
00:26:53
Speaker
But those... race dynamics had an end, right? We had Kennedy's speech about putting a man on the moon by the end of the decade. Like that was kind of the line in the sand.
00:27:05
Speaker
I, what can we say about the race with China? It's not like there's no, what's the finish line? What's the end stage? Yeah. um So I guess i you have talked with people who are sort of that seems to be the political justification for deregulation for allowing the licenses to go forward.
00:27:25
Speaker
Just want to get your sense of, you know, how much work is that? Is that justification really doing here? I think what you said about the smaller models is really interesting. Just to point to that first, I spoke to um sas so Sasha Luciani was previously the climate lead at Hugging Face. She just started. Big of hers.
00:27:47
Speaker
She just started um a consulting firm of her own. She left hugging face and she'll be working on sort of helping enterprise companies understand that, you know, in many cases they can choose the type of AI they'd like to use. Maybe not every use case needs to go to a huge LLM that requires a huge data center sort of educating and, and you know, de-risking themselves from these issues.
00:28:13
Speaker
from these big data centers. We also don't know, as you say, regarding the space race or or the race with China, we don't know how the technology itself is going to change. So my prediction is that there are many, many people I've spoken to who are working on other advances that would make ah models more efficient, that would make chips more efficient. And I do wonder whether these mega AI data centers won't become as, won't be as necessary in a few years as they are now. And I know that some communities are concerned about that, that you could end up with sort of these hulking ah frames of of data centers that in this case are not really good for that much else. Like you could have another type of data center there, maybe break it up into parts. As far as the AI race with China, i don't think there is an end game there. I feel like it's more like...
00:29:08
Speaker
um you know, sort of an ongoing one-upsmanship that, you know, some people think could be very dangerous. Some people think, you know, we'll just end in some sort of stalemate, you know, based on negotiations like Trump just had in China this week. um But yeah, I feel like having that as the, you know, as as the, you know, one-upsmanship is sort of,
00:29:37
Speaker
probably not the greatest measurement of how successful we are in this in this realm. um I think that... you know, these communities in particular would like to understand more how it benefits them. And that's where I think better communication from companies needs to happen. I mean, the problem is, is that they're promoting use cases like for knowledge work, which may or may not seem that important to the average person. And those use cases will hopefully translate into more, know,
00:30:13
Speaker
I would say really useful use cases like curing cancer, for example. yes I think, I mean, my casual observation is that despite the quote-unquote race with China, I feel like we might actually be running two different races because, you know, they a few months back, there was a lot of laughter and ridicule about, um I think it was Baidu robo-taxis kind of shut down on a Chinese highway because of a software update.
00:30:41
Speaker
And in the West, it was easy to be like, aha, that's why we drive our cars. But I was like, You're hearing about this like a plane crash because it's rare. But can we also point out that it's a country that has like 40,000 robo taxis like on a highway, not Waymo's circulating limited in a city.
00:30:59
Speaker
ah The New York Times had some articles about consumer products using AI like... ah parents being able to wear this mask that they can speak in Chinese and it translates in real time to English to teach their kids English. And I i think we have a lot of technological innovation, but I think we're struggling to get it into the economy in ways that are more meaningful than, say, chatbot girlfriends or, like, it can write my emails for me.
00:31:28
Speaker
Well, I will say that I think part of the problem is that It hasn't, not sure when it will, trickle down, you know, some of the more advanced use cases to the general public. Because I see, you know, I see folks talking about really interesting use cases for AI. But, you know, maybe they're really expensive. Maybe they're things only very large companies could do or very...
00:31:55
Speaker
ah folks who are very technical still can do. You know, like the question is, how much will this democratize? Because for example, ChatGPT has several different payment tiers. If you're using the free tier and you're not paying for a Pro or a Max, you're not paying for for a codex or, ah you know, in Claude, you're not paying for Claude code.
00:32:19
Speaker
um Maybe you're not really understanding how much these models can do. um But maybe you you never will because you're not, you know, you can't afford to pay for it or you're not going to.
00:32:32
Speaker
Yeah, I mean, like if you're just using it as a therapist versus trying to build an application you want to take to market, I think you're your concept of the platform are two very different things depending on purpose. So I think you're bang on with that.
00:32:45
Speaker
Um, in general, I even got close out with this. like You've watched the story go from, as you mentioned, a niche infrastructure beat to something showing up now in governor's races and reshaping rural America. This is now going to the midterm. I think one of the most important issues is going to be debated over and over again across the board nationally.
00:33:04
Speaker
What is the story you think the rest of the press is still missing out on when it comes to this data data center, um, i don't want to call it problem, this data data center conundrum that needs to be solved sooner than later before the environment's completely destroyed?

Political and Technological Speculations

00:33:23
Speaker
For me, I think it's really... really understanding the kind of backlash that I think is being fomented here. um I was just finishing up my story about a project outside of Waco, and the state representative was saying to me that um a year ago, these data centers were not on any ah Texas politicians radar. Now it's part of their platform that they're putting forth for November. They see how much this is bubbling up, but I'm not sure how much the tech world, I'm not sure how much the general public really understands how this is bubbling up. And I fear actually
00:34:06
Speaker
ah very much for some of these conspiracy theories that are coming about. Like as, you know, we can criticize data centers, we can be very concerned about their impact, you know, real impact on the environment or health or water or anything like that, but they're not mass surveillance centers. They're not population control buildings. and I do worry about ah that becoming part of the,
00:34:36
Speaker
the take that takes us into November, I think that'll be pretty dangerous. Yeah, it's a good point. I mean, in light of, you know, the two attacks on ah Sam Altman's house and a few others, even attacking, i think, local city council members who have um voted in favor of data centers. Yes. I i think,
00:34:59
Speaker
If people feel disenfranchised, as we said at the beginning of this show, like they're saying no and still it proceeds. You can only sort of squeeze people, but so long before conspiracies take place. And if conspiracies take hold, sometimes violence is quick to follow. Yeah.
00:35:16
Speaker
Yeah. the Defending democracy, I think the perception of democracy and and the people's perception that there is no democracy and there is no freedom of will for their jurisdictions and for what impacts their lives.
00:35:31
Speaker
I think that leads to the foundation of civil unrest, which is kind of what you're referring to. But it also just can be in a very local community with local part-time officials. These things can be based on real laws. I mean, zoning laws are real. Most of us don't know what they are typically in our community until, say, we decide to build an addition to our house. um But, ah you know, that's one of the interesting things that I've found out that, like, you know,
00:36:07
Speaker
It's not true that you can, that a town can just object wholeheartedly to any project coming to their town. There are laws and there are rules and and every state typically is different.
00:36:20
Speaker
so And also interesting, one of the follow-ups is in a similar town in Festus, Missouri that Wall Street Journal was in. Very, very loud town hall meetings around a data center. And I think they also gathered enough petitions after the vote, which was in favor of it, to recall all of the members who voted yes. Anyway, I'm just saying that people are trying to find different mechanisms to, I guess, make their voices. Absolutely. In Saline Township now, afterward, they're looking to recall three of the
00:36:55
Speaker
ah City Council members. um And in the Texas town that I'm finishing up my story, the same there, there, there, there's a new City Council member who is opposed to the data center. And that's why she was elected because she is is in opposition. So again, there's just a lot of these.
00:37:16
Speaker
different things that are happening. And I think the main issue is that this isn't just one town. This isn't just yeah a few towns. This is happening in hundreds of towns across the country.
00:37:27
Speaker
um Last note, you so mentioned before we started recording that this was a series. um As the series comes to a close, where else is your journalistic gaze headed in terms of the AI industry?
00:37:43
Speaker
I think there's so much more to look at besides, you know, although I'll certainly want to follow up with these communities that I've followed and certainly there'll be new projects to look at. But I also think I'm really interested in sort of what ways ah developers are looking at to make these projects. If they really are coming, um you know, how are they going to be more efficient? Do they need to be this big forever? could Are there ways to um co-locate them? So there's just one smaller building in one place connected to another in another place. I really do feel like there'll be more advances in technology. i mean, some people, i wrote a piece about ah putting data centers in space, which a lot of people consider very unlikely within the last ten next 10 years. But there are many people working on the problem. And i can imagine that in 20, 30 years, you'll have it if needed.
00:38:46
Speaker
Sharon, thank you very much for your reporting. And most importantly, as we mentioned before recording, physically going to these tiny towns and and not just sitting at the desk. So thank you again for your time and your attention. We really appreciate Outstanding work. Thank you, Sharon. Thank you so much.
00:39:06
Speaker
All right, questions to take forward. We usually pose questions to the listeners. I think for listeners, I would just tell you to start digging in a little bit more behind the headlines and announcements of data center build outs. But I actually would pose a question to the founders and the AI developers, which is how will technology that you want to bring to bear touch on real human problems? The frontier labs are ahead. They're going to do their thing. They're going to race for their so-called AGI or build their machine gun or whatever.
00:39:42
Speaker
But there are plenty of applications there for machine learning, deep learning that could really, really help people's lives, whether it's in biotech, whether it's technology.
00:39:53
Speaker
detection. you know, I try to highlight these wins on on LinkedIn, but i I would put that to the founders. Like, think about how you can bring the magic of AI to bear on people's actual lives.
00:40:07
Speaker
Yeah, for my question, sort of take it to the common listener. I would say, you know, I've long known the real power in politics at the municipal level. Do you know your zoning laws? Do you know what the local constitutions are?
00:40:21
Speaker
do you know what power you actually have as citizens? Because I think if you're going to effectively fight it, sure, protests are great. Sure, showing up at the rallies are fun. Sure, yelling at your city council is cool.
00:40:33
Speaker
That's not where the work gets done. That's not where the work is. The work is understanding policy. Do you understand the policy and you understand the power that you have within it? And that's the question I'd pose to you.
00:40:44
Speaker
That's awesome. I will raise this quote from one of Sharon's articles. So she interviewed this local woman. She was a Marine in Afghanistan in 2012, moved to Saline Township in Michigan to raise her family. This thing's going up across her lot.
00:41:01
Speaker
So I think her name Catherine Halshalter. And she says, quote, I've had to learn more about ordinances and state law and zoning than I ever thought I would want to. But now I realize how important these nitty gritty, seemingly boring things are. They can really upend your whole community, end quote.
00:41:19
Speaker
All right, we'll leave it there. Take it forward. We will see you next week.
00:41:27
Speaker
If you like this conversation, share it with friends and subscribe wherever you get your podcasts for a weekly ballistic payload of snark, insights, and laughs. New episodes of Bare Knuckles and Brass Tacks drop every Monday.
00:41:40
Speaker
If you're already subscribed, thank you for your support and your swagger. Please consider leaving a rating or a review. It helps others find the show. We'll catch you next week, but until then, stay real.