AI Data Centers Have Hidden Costs

[00:00:01]AI data centers have hidden costs.

[00:00:04]I'm Nick Espinosa, your chief security fanatic, and let's dive in. Now, for years the conversation around data centers has focused on a couple of different topics, right? Energy consumption being a primary one, not to mention water usage, carbon emissions, and all of that.

[00:00:19]But there's another issue that's emerging, and it really underscores that we don't know all of the unforeseen costs of an AI data center. And that new issue is noise pollution. Because according to recent reporting by the New York Times, and I believe some others, residents living near large data center campuses are increasingly complaining about the constant noise being generated by the AI data center. We're talking cooling systems, backup generators, fans, chillers, other infrastructure basically that is essentially required to keep what is a modern computing facility running. So, in some communities, residents are now describing a persistent low-frequency hum that operates 24/7/365 and is impacting their quality of life.

[00:01:04]Now, what makes this important is that I think it really highlights something that we forget. And that is that the cloud is not really conceptual. It's not floating around in the sky, for example.

[00:01:16]The cloud is somebody else's computer, and in this case it's actual physical buildings in communities that are now generating a ton of electric electrical use, water use, and noise pollution. And so, AI basically the demand we know is accelerating. And so, by virtue of that, data center construction is expanding rapidly, not just in North America, but also in Europe and Asia as well. And basically the communities that are once welcoming, or originally welcoming, these facilities are now really starting to look at the local costs. Because part of the sale of this is it's going to bring jobs to your local community. Well, outside of the construction jobs to build and eventually uh lower population of construction workers to maintain the facility. There's not a lot of people that work in a data center. It's just a whole bunch of computers in a giant building, right? But, I want to talk about why this matters here because most people interact with data centers day-to-day, and we don't really even think about it or consider it. But, every time we use like a Google or Instagram, Netflix, uh chat GPT, Anthropic, whatever, our email through Microsoft 365 for business, cloud storage, you name it, you name it, you name it. All of this is massive physical infrastructure all across wherever you are, right? Now, historically, data centers had been relatively small, but AI is absolutely changing that because the newest the newest AI-focused facilities are requiring a ton of electrical, cooling, uh water, not to mention land, and just physical infrastructure as well. And so, the result seems to be that communities are beginning to feel the impact that they really didn't think about, anticipate, and those are big, big issues. And it's becoming a global issue, and let me tell you why.

[00:03:05]Because the world is entering what some analysts call essentially an AI arms race. Specifically, the infrastructure that AI needs. And so, governments and corporate corporations are moving and competing as fast as they can to build AI training clusters, hyperscale cloud facilities, semiconductor manufacturing, that is a huge problem seeing as how Taiwan and South Korea have locked down like 98% of that field. Uh Taiwan being 90%, that's not good on the China side.

[00:03:32]Uh you know, so we've got manufacturing capacity that needs to be expanded.

[00:03:37]Advanced computing, uh you know, it is is the name of the game. On top of it, we need infrastructure as well. It's a huge, huge thing. And the assumption is that more compute has always equaled more innovation. But, the problem is that there are physical limitations to that concept. The digital world might be virtual, right? But the infrastructure supporting it is very real. And so, let's talk about why we need to put limits on data center expansion. And I think this is oftentimes just very overlooked when we are discussing this longitudinal issue because the question isn't whether or not data centers are important. They obviously are. We need data centers. You would not be listening or watching this without having a local data center pumping this out through a content delivery network for me. In this case, probably Google's data centers cuz this is on YouTube or SoundCloud, which I believe is on AWS. The real question here, though, is whether unlimited growth is sustainable because we are looking at a period right now where it seems like unlimited growth is what they are trying to achieve with AI data centers. I mean, physical resources are not unlimited, right? We have a finite amount of of things like water, electricity, etc., etc. that can go around or we need to build more infrastructure. But we've seen water issues alone, at least here in the United States. So, we have a finite amount of resources. Go look at Thomas Malthus, you know, past economist for for his Malthusian theory on on the earth, which quite frankly kind of being revisited today. On top of it, we know that local communities are bearing the costs. Residents near data centers, on top of the noise, you know, they're seeing their energy bills rise.

[00:05:18]They're seeing water bills rise cuz the water consumption is going up, not to mention just changes in land use, environmental concerns with all the output from the massive cooling systems these things require. And so, the profits for things like AI data centers could be a global phenomena just given what AI companies are. But the consequences are being felt at a very local and you know, very at grassroots level here, right? And if you think about it, the demand for artificial intelligence just doesn't have a natural stopping point. You know, I was thinking about this, you know, that that this is going to be a huge problem because one of the biggest challenges that AI companies, you know, are incentivized to to to do is continuously expand, right? There's no mechanism, you know, in an open market that says, "Yeah, this is enough. Let's shut it down." There There's nothing like that here. And so, infrastructure decisions also can outlast governments as well. Because a data center that, you know, we build and go online tomorrow could operate for literally decades.

[00:06:20]What gets upgraded is the computing system inside the building itself. And so, future communities inherit those consequences. In other words, won't somebody think of the children?

[00:06:30]You know? So, long-term planning is beyond critical to ensure that that we have a balance here. And outside of that, there are other opportunity costs as well, right? So, like every megawatt allocated to a data center is a megawatt that can't be used elsewhere, right? So, if we are taking land to develop a data center, you know, we can't basically use that land for anything else. And so, it's not about whether AI is valuable or not. It clearly is. I I see the positive effects in my life, at least my business life, as I'm using it to cut down on things like research time. I don't have to flip through 20 pages of Google results when it will give me the result I'm looking for immediately and all of that. But, the issue here is whether we are properly evaluating the tradeoffs for society as a whole. And I think that's something that we are absolutely missing. And we have here the ability to avoid past mistakes. We just have to look reflexively a little bit, right?

[00:07:28]And just reflect on where we've been because the internet, social media, you know, not to mention large-scale data collection has also followed a similar pattern, right? It's deployed very rapidly. It has massive growth. There's unintended consequences, the governance sometimes doesn't come, at least in the United States, and we have that ability.

[00:07:48]I spoke about this the other day as well, to reverse those kinds of issues.

[00:07:53]And so, it made me think, like, what should happen then before we start to we start to really expand or continue the expansion that we are seeing. I think we need to have community impact assessments, and then communities need to understand, like, noise, water, energy, environmental, all of these kinds of impacts are going to direct them, and they're going to direct them impact them directly. That's a huge thing, and I think those need to be looked at before approval is granted for this. And and there should be limits on the uses that a data center can have, right? A government, whether it's local all the way to federal, could set a clear expectation around how much water, energy, and land is consumed to make the data center the data center, right? Just don't allow unlimited expansion. We saw one in in Utah where they had to scale it back after public backlash, but this thing was going to be something like a thousand football yards long or something. It was absolutely massive in the middle of nowhere, but the amount of power that thing would take is is a huge issue. And so, we need to encourage efficiency instead of just pure growth, right? And we should, for the record, incentivize the industry to do this. You know, more efficient computing systems, better cooling, you know, all of that, lower energy consumption, all of that should just be there instead of just building something bigger and bigger and bigger and bigger.

[00:09:14]And so, the communities I think need meaningful input on this. You know, if you're living near infrastructure, you should have a voice because this affects you directly, right? And public often discovers, you know, a lot of these things after construction has begun, or they don't question what's going up in their neighborhood. Nobody attends town council meetings, you know, like, per capita. Every time I've been to a town council meeting based on my job because I'm let's say I'm presenting to a city council because I'm working with a city council, there's like three people in the room beyond the actual counselors.

[00:09:45]We have to start getting involved in these kinds of things because I think there is a bigger lesson here as well.

[00:09:51]This is something this is about something rather larger. It's not just about data centers because for years, literally years, we've just assumed that these digital technologies were kind of weightless and limitless. People don't really think about what the cloud actually is. They just know they can save to iCloud or Google Drive or whatever, not really thinking about that. But AI, I think is really kind of revealing this assumption was never true, right? That that that people can see these things tangibly going up in their neighborhoods. And so every digital service ultimately depends on actual physical infrastructure, which means actual physical resources, which means it's impacting actual physical communities. And so the question here isn't whether we need data centers. We obviously obviously do, right? Again, you wouldn't be listening or watching me without a data center. You we wouldn't be able to to use the internet in the way we know without data centers. The question though is whether we continue treating computing power as like a limitless good, right? Or whether we actually start recognizing that every digital breakthrough that we are going to achieve beyond AI even carries real-world costs that someone somewhere has to bear.

[00:11:07]And that's something that we can't take for granted. So, please do your homework. And if if there is a data center proposed data center coming into your area, really push your town councilor or whoever's in charge for impact assessments. And it's not just on water, energy, and land, it's also noise as well. And I know there'll be something else we're not thinking about in the future. So, best of luck to us all.

[00:11:29]AI data centers are coming, but it's our job to make sure that we are not letting them go unlimitless.

[00:11:37]And please like, share, follow me here on Facebook and Twitter at Nick A E S P, courtesy of your local data center. Same when you subscribe to YouTube as well.

[00:11:44]There should be limits. And as always, stay safe, stay online, and please, please, please, I don't want to say private, informed, and secure.

[00:11:52]Take care.