Latest / Computer Says Maybe / Net0++: Data Center Sprawl | NEW Research from The Maybe
Transcript
- Alix: [00:00:00] Hey there. Welcome to Computer says maybe this is your host, Alex Dunn. This episode is actually about research we've done and I'm really excited to have two researchers, one Prathm Juneja, who you've heard before, the other Chris Cameron, and they were two of five authors of a report that we are releasing on data center development in five different countries. Looking specifically at case studies. Where we think we can better understand the dynamics of how local policy makers are consulting communities, how process unfolds when a big tech company comes to a city near you and says, Hey, we wanna build a data center 'cause we're just gonna help you create jobs and it's gonna be great and you're gonna be part of the innovation economy. Isn't that exciting? So we looked at five different case studies to kind of see how that plays out. When big tech companies have a ton of lobbying power, a lot of information, a lot of narrative, a lot of sort of wind in their sails of being able to tell a story [00:01:00] that may or may not be true, um, and communities and organizers and journalists and people that are interested in really knowing, do we want this thing in our backyard? When those people get activated, what happens? So with that, I dig into this report, why we wrote it, what we took away from it. What you might be able to take away from it. So let's dig into it with Chris Cameron and Prathm Juneja. So around the world, data centers are being built super quickly with very little regard for what people in the area around the data centers actually want in terms of how resources are allocated, in terms of how the process goes, in terms of a company. Approaching a local government and saying, we wanna build a data center in your backyard. It feels like the ways that this was happening was all [00:02:00] bad, but how it was bad changed based on context. So we wanted to do a deep dive in several different countries, which we'll talk a little bit more in a bit about how we, I. Picked the countries just to understand more deeply about the mechanics, the politics, the process, the resistance to these data centers. So we assembled a team of researchers who could take on this project from these different vantage points and work in these different case studies. And I am pleased to have two of them here to talk about what they found in the research. So, do you two wanna introduce yourselves? I'm Chris Cameron, Prathm: and I'm Prathm Juneja. Alix: So you guys were a part of a team of five people, six people. If we count the people helping with design and editing of the report, I know that most of the people listening to this podcast know what data centers are, but do either of you want to feel the question of like, what are we talking about when we talk about data centers? Chris: So data centers are these large facilities that house infrastructure to support and power digital applications. They're used for ai, they're used for holding and [00:03:00] storing data. So anything that's on your phone. So in general, we're kind of interacting with data centers every day. Prathm: Yeah, they're, they're big computers that are not located with you, but elsewhere, and you access them via the internet most of the time. Alix: And when we were started this report, do we wanna talk a little bit about how we decided to structure it in terms of looking at not just a single data center, but sort of thinking about it from different contexts and how we approached it? Chris: Yeah, so we, we are four researchers and we were living in different countries and in different contexts. So we were thinking about how to write a report that could represent multiple regions and communities across the globe. And then we started researching into different data center regions, and we found that the biggest issue with data centers is that there's no data on data centers. Alix: Uh. That's the irony. It's the irony. Do you wanna give us like the overview of the thing in terms of the countries we worked in and then maybe [00:04:00] Chris you could talk a little bit about the kind of angles we took when thinking about each of the case studies? Prathm: Yeah, absolutely. So we had five case studies. We did Santiago, Chile, peculiar, Missouri Z Wday in the Netherlands, ero, Mexico, and Rohani South Africa, or in nearby Johannesburg, South Africa. Alix: Cool. And then Chris, do you wanna talk a little bit about like what we were looking at maybe across all of them, even if particular ones took us in different directions? Chris: So I guess we looked at like all facets of the data center. So we looked from the initial planning stage to breaking ground, to the construction, and then finally the outcome and how the community was reacting or living with it, or if. The data center was canceled or they managed to stop the building process. So we also looked at how government agencies interacted with data center companies, and if there were some kind of incentives that were given to data center companies like tax exemptions or for [00:05:00] example, in ero, Mexico, the government actually gave away free land to Cloud HQ to build a data center. Also, things that were unearthed. During investigations by journalists, which shows how important their work was. During the research and writing of this report, we also looked at how communities received data centers and what were their strategies to resist, and then the outcomes kind of in how policy might have changed or how local governments were responding to the data centers. Ultimately, we wanted to identify what resources are needed right now for this work, for activists, for advocates and researchers, and how to strengthen community advocacy. Alix: Yeah, and I feel like each of you took on different contexts and then in the process of that and ended up kind of going in slightly. Different directions. Um, do you guys wanna talk a little bit about the kind of themes across the case [00:06:00] studies pro, maybe starting with South Africa and some of the challenges you had and even like finding out what was going on? Prathm: Yeah, absolutely. So I, I wrote the case study about Equinox's JN one, which is a relatively new data center that's located quite close to Johannesburg. It's like right across the street from Johannesburg's Airport. And this is a big data center. It's $160 million. Co-location data center, which by that we mean it's not specific to one company. Equinix builds this data center and then they lease out the space to companies that want to use it in a variety of ways, rather than, say, Microsoft having its own data center that's only used for Microsoft products or something like that. The simple story for Equinix is JN One is that we don't know much about it. There doesn't appear to have been a lot of community resistance against it, and it was built with a lot of support from local and international governments. I mean, a bunch of local officials in South Africa spoke at its launch event, but so did the US Ambassador 'cause Equinex is [00:07:00] a US-based company. The key kind of context here is that Gatting and the Johannesburg region have been dealing with water shortages and load shedding and power shortages for a very long time now. It's one of the key socioeconomic issues that in many ways is tied to the legacy of apartheid. Water access rights, or individual water access rights are owned, I think like 98% by white people in South Africa. So there's drastic post-colonial implications of this, and data centers, as we know, have a tendency to use a lot of electricity and water, especially air cooled data centers use a lot more water than most water cooling systems, which might be unintuitive to some people. And we just get all of these promises about how this data center is gonna create a bunch of jobs for the region and how it's gonna be completely environmentally sustainable. I mean, there's a quote that's something like, this data center created a lot of jobs for women and children, and it's like, what? How, in what context did that happen? Please give us some evidence of [00:08:00] that. How much energy is it using? How much water is it using? How is it affecting the lives of the people in nearby Johannesburg, but also in the townships that are near the data center itself? I also spoke to a researcher, Samantha, who had. Spent some time interviewing people near and around the site who basically found in a lot of her research, interviewing these folks that they hadn't even heard of the construction of the data center, but they were deeply skeptical of its ability to actually improve their lives. And I think that is part of the transparency story here. It's not just that these companies have a tendency to withhold information, it's that they don't even, in many cases, view it as their duty to inform. That this is even happening. It's not just like you start with a fake community consultation, it's that they didn't even do a community consultation for the people who are most likely to be impacted. So the story with the JN one data center is that I think we've subtitled it something like a success, but for whom, insofar as it is a. Drastically successful project for Equinix and for the people in the digital community and South African government [00:09:00] who supported the construction of this data center. And it will probably bring in a lot of corporate money, but it carries with it this. Pretty serious lack of transparency on its potential impacts, which we just kind of have to guess are bad in the absence of information and based on what we know. And it's going to affect people who are already being affected by some of the most severe harms of water shortages and electricity shortages in a way that's really disappointing and I think tells the start of a story that we are gonna continuously see of Western Data Center and tech companies. Building data centers in the global south that are probably gonna result in the exploitation of those folks. This continuation of technical colonialism that we've seen in a variety of forms over the last few decades that are gonna continue to see from data centers. Alix: Chris, do you wanna talk a little bit about the trends you saw across the two case studies you did? I Chris: did two very different case studies. So I looked at in the [00:10:00] Netherlands and I also looked at, um, Mexico. In the Netherlands, it was quite interesting how. The community organized so efficiently, and maybe this is part of Dutch culture, where they were so ready and prepared to create a Facebook group. They petitioned all of the residents of the town were aware of what was happening and all of the information was shared between them. And they were using Meta's product, Facebook against themselves. 'cause what was happening was Meta had plans to build a hyperscale data center. On farming land. So they were pushing out farms and this is when the community started getting involved because it was a community centered harm, the residents of vol. They also connected with and used case evidence from investigative journalists at uh, local Dutch newspaper. And they were debunking basically the data [00:11:00] center company. So Meta's claims that they were gonna use the residual heat. From the data center to heat the homes of, they used architects to also identify that it was meta. 'cause Another covert tactic that Meta was using was they were using a shell company to initially buy the land and they call themselves Operation Tulip. This was an issue because there was, you know, a lack of accountability on me side, but also this lack of communication from the local government to their residents. And so I think that also stirred up a lot of anger in the community. 'cause they, they wanted to know what was happening and there was no information. What happened was. Their case evidence and the organizing led to them influencing the local election, and they actually pushed out the alderman who wanted to bring in the [00:12:00] data center. And so they elected a new alderman who ended up being a data center abolitionist, and he canceled the whole project. But what happened after that was that meta moved the hyperscale data center to Spain, and so they ended up. Building pretty much the same plan. They called it Operation Zza, and they built it in Spain. Alix: Okay, so basically if we take the Netherlands case, it's essentially like, uh, small business wants to build a data center in your backyard. And actually that small business is meta obfuscating. The fact that it's meta, and if people had known it was meta, they probably would have approached the whole thing very differently. Also, like so sketchy to name it Operation Tulip, the fact that we're talking about generative AI boom, which is a speculative market play. And tulip, I mean, I presume you guys know about like the tulips. Um, yeah, yeah, yeah. Okay. Is that the, were they making a joke about how hype is creating pyramid scheme, like structures in the global economy? Chris: [00:13:00] Because that's like, I think it was too, that's too much for them. I think they were trying to be cutesy. I think they were like, it's Amsterdam. We love it. We love Exactly, exactly that. Completely agree with Chris. That's what it is. That's what I think it's, but yeah. Um. So that case study was so different from the case study in Quero, and I think it also, from the research that I was doing it and from the conversations I was having, there was quite a lot of Nimbyism that came from Alde and the residents Nimbyism, like not in my backyard. The Nimbyism was also seen a lot in some other case studies we were looking into, such as in peculiar Missouri. Also, initially we were looking at London, some data centers there in Ham. You were looking at those, but we found they were a bit too similar. And we also wanted to include some case studies with different situations, so different statuses of data centers, whether they were completed or canceled. And so [00:14:00] we looked at in Mexico, and this was actually a project that I was investigating initially with Emma Press, who also is an author on this report. The difference that I really saw with the community resistance in ZedE versus the community resistance in Mexico came down to how their government structure was and how local government was interacting with their citizens. It does relate to Nimbyism and how there is like this privilege to have the fundamental rights to protest and feel safe to do so. And in Mexico these rights should be available, but they're not. And the majority of environmental harms are occurring in places where there are no fundamental rights to protest or there is a lot of violence if you are an environmental land defender or journalist. And so it was quite different when I was speaking to journalists or community organizers in Mexico because a lot of their [00:15:00] stories were about state violence when it came to protests about how they were. Trying to protest their constitutional rights so that they could have access to water. There was a through line that I found between data centers, water justice, and indigenous land sovereignty. And so the case study itself was a much bigger story than just a data center coming into a community and disrupting the area. One of the recommendations that we had coming off of this case study in particular was having this interconnectedness of community organizations. So even though there might be a community org that's focused on water justice or environmental justice, still including them in the data center conversation and also finding efforts to fund, convening to bring together these communities as well. Alix: Any reflections on sort of how we structured it? You're in the middle of doing a PhD. I imagine that the research process was a bit different than the [00:16:00] research you're doing, uh, in your, in your daily life, but any reflections on the strategy of structuring the report and the way we did? Prathm: Yeah, absolutely. I mean, so the goal with the report was to cover. The impacts of data centers on the communities that they are built near or within or on top of. And then also how those communities have resisted successfully or unsuccessfully or at all the data centers themselves, the construction of, um, operation of them. And so what we wanted, as Chris noted, was to try and do this across a variety of geographies and contexts because none of this is gonna be static in any specific region. And so we spend a lot of time trying to think about where. Data centers are being built and you know, what communities are being harmed the most and what stories we could tell. And as Chris noted, finding the information necessary for something like that was deeply difficult. Not just in the global south or in Asia and Africa, where it was absolutely the most difficult to find reporting or transparency or information. But in the us, in the uk, et [00:17:00] cetera, like finding proper information about. Who was building these data centers, why they were being built, what they were doing to the communities, how communities resisted, et cetera, was nearly impossible for most of the centers that we had in our initial shortlist, which is how we resulted in the fight that we have this amount of information for. Alix: I mean, it's a bit of a, maybe a spoiler, but do we have a sense of why it was so hard to find information? I mean, I presume it's intentional obfuscation by companies who. Wanna hide what they're doing. 'cause if it was clear, people would be more likely to be like absolutely not. Is that the general, Chris: yeah. So I think the issue too was that it kind of fell to the individual. So it was the community activists or it was the investigative journalist. Who felt like they were trespassing when it came to trying to access information or trying to use FOIA or look at, uh, country's national utilities files. But really this was more of like a covert tactic from data center companies to hide their information. So they couldn't [00:18:00] really share what they were doing in regards to the environmental impacts or the water and energy use. And we found that this was a shared sentiment across our cases. Prathm: I mean, a lack of transparency comes from a variety of sources. Obviously the primary one, as you both have noted, is why would a company share this information if they're not forced to, if it's going to harm their cause? Like obviously the financial incentive there is simple, but also just, we talked to researchers and people trying to get some of this information and environmental activists who express to us that there are for many of the people working in resource constrained areas, more pressing issues than the data centers, right? And so there's also just a lack of resources. Dedicated to trying to find the information. So it's a mix of the information is not readily available for variety of incentives, but also the amount of time and effort in some cases required to get the information past those obfuscations is also quite difficult and there's some sort of prioritization that needs to happen for the limited amount of people who are working on the subject. Alix: And also probably top down [00:19:00] standard setting about what companies have to disclose proactively. Because I feel like the, and this is actually something Karen, how brought up in our conversation last summer, basically, by definition now, any reporting on this is investigative in nature, which has to change because if it doesn't, it's not gonna be sufficiently transparent for communities to actually participate in any kind of decision making. Um, do we wanna talk a little bit about the motivation for the report? Like I'm happy to share. Perspective on that, but I'm curious when you guys were thinking about the structure of it, like basically the current state of actions happening to help with this. Yeah, and like what we see the role, we see the report playing. Chris: Yeah. I mean, not to be depressing about it, but we're kind of like at the peak of unprecedented world ending conditions. Uh, the climate collapse, the pace is increasing. Our governments are being taken over by AI driven techno imperialists, and I. See the global health of the world kind of degrading from increased pollution and worsening air qualities.[00:20:00] So there's like a lot of hopelessness, I think. But one of the things that I learned from working with communities experiencing environmental racism is that action is an antidote to grief. And so I think a lot of the motivation is having agency and controlling your future. And I spoke to a lot of the communities resisting, and they felt that it was important for them to take it on because it was their environment, it was their life, and I. They were the ones who were living next to the data centers. Prathm: The one thing I'll, I'll add to that is we spend a lot of our time in mainstream public discourse, taking data center construction as kind of an inevitability, and the way that it's happening is just inevitable. And it's not a thing that we think about because the main story is the AI race and. The vast growth and rapid growth of ai, and everyone's taking data centers kind of for granted there. But as a result, we might talk about the harmful impacts of ai, but we're not talking about the harmful impacts of data center construction. And it is a key part of this story that often gets [00:21:00] neglected. And so part of the motivation here is we want to shine a light on something that. There are people who are dedicating their lives, careers and like have to dedicate their lives and careers to trying to oppose or reform, and they just aren't getting enough attention about it, which goes to the transparency and information availability problem as well. Alix: Yeah. I also feel like there's a depth in physicality to these topics that a lot of people don't know about. If you look at this problem globally or you look at this situation globally, it's very clearly. A giant terrifying trend. Um, but if you look at it locally, maybe you only experience one time where there's a massive data center facility being built. Global actors like companies have this advantage of being able to act basically like hurt a community and then move to the next community who doesn't know what's coming. And I think there's value in joining up those stories in a way that makes it clear this bigger trend, and also makes it easier to articulate why these companies should be held accountable and in what. Ways, because if you look at a single [00:22:00] municipality, that story, if it's disconnected from other stories, feels much less problematic and you don't see the kind of bigger pattern. Prathm: Yeah, I'm, and one of the main goals of the report when we first set out was can we find examples of resistance that worked to reform or oppose a data center in such a way that it was positive for a community? And then can we find ways to generalize that approach so everyone can use it? And it's like, obviously we did not come up with a list of here's what to do if a bad data center is being built in your neighborhood and how you can resolve it. But it would've been nice to, because. Again. Yeah. As you mentioned, the companies have the real advantage here. These are disconnected networks in completely different contexts, often, and they can use the same tactics to advance the construction of a bad data center in a neighborhood, but people might not be able to use the same tactics to oppose them Chris: data center companies learning from each other. We found that in our research there was one called data center dynamics.com, and it would also tell data center companies how [00:23:00] to interact with resisting communities. I. So it was an idea for us to then be able to. Write at least some ideas of how to move forward, some structure to support communities in connecting with each other and to network that and to be able to share resistance strategies. So that's part of the report as well as we looked at the different strategies of community resistance. In each of the cases. Alix: Yeah. I will say, I mean, picking up on something Ham said, we did act in good faith to think are there conditions within which building a data center is an appropriate thing? Like I wouldn't say I'm not a Luddite, I think I am a Luddite in terms of thinking about how new technologies. Open up opportunity for increased exploitation and concentration of power and that that's a problem. I also think that like having computational resources is something that as a society we broadly probably want. So there could be situations where it's [00:24:00] appropriate to build a data center that can power those computational resources. So we kind of went into this project being like, there must be. An example of a data center that was built in a place that had a sufficiently capable energy grid, had enough renewables on it that like having a data center suck a bunch of energy, wasn't gonna result in more burning of coal. Chris: I would say I am a data center abolitionist at this point from all of the research that I've been doing, and the only good example I could really find where a data center has served the community was during a hurricane in Texas. Uh, data center has enough power and water to sustain, I guess an an extreme weather event. So communities actually hid in the data center, and that was its purpose was more of shelter. Um, but yeah, for now, I, I agree with your themes, Alex. Like, I really see. The expansion of data centers, part of this creed and movement to gather as [00:25:00] much data storage as possible to supply AI and the differences in a traditional data center. The traditional data center that's maybe used more for data storage. Uses like 10 kilowatts per rack. Whereas an AI data center uses 60 kilowatts per rack of energy. So the usage and energy storage is much higher, the water usage is much higher. You just see this exponential increase with ai. Prathm: Alright, I'll push back. I'll do it. Um. I agree with both of your takes on this as the current situation, but I think that we can have a positive vision here. So yeah, we have some examples in the report of very small scale attempts to make this better. I can't point to a major data center company right now and be like, these guys are always building really good community, helping data centers. Like absolutely not. That is not happening, and as is true with basically any industrialization under capitalism. [00:26:00] If you don't put limits on companies, they're not going to immediately start caring about people for no reason. Like I, I don't think data centers are going to be unique in that regard. And because there's so many additional harms in the construction of data centers, from where you place them to the environmental impacts, to the job impacts, to the noise traffic, et cetera, the harms can be quite severe. But I do think that we can have a positive vision here. I believe that there are going to be, and currently are good reasons to build AI systems. Not all of them are good, but I think that AI can be useful in a variety of ways in society and will probably improve our lives in some ways if we do it well and carefully. And so in order for that to happen, I think we need data centers. And I think that we can build data centers in ways that are positive, that are sustainable, that create jobs for a local economy that don't destroy. Local water and electricity resources. I don't think this is actually happening at scale at all right now, but I think it is possible, and I think because we have to deal with the cards we're dealt, I don't think we're successfully going to get Microsoft to stop building data centers tomorrow and Equinix to stop building data [00:27:00] centers tomorrow. So what does it look like? As our recommendations really suggest in the report to fund the kind of work necessary to help improve some of these outcomes while also stopping the worst ones from occurring, obviously Alix: get trolled from the center. Um, yeah, there Prathm: you go. Alix: Uh, I, uh, I mean, I think you're right in theory, like I think there's a, the excesses of this industry to me. I wouldn't describe myself as an abolitionist, but I also think that abolition as an anchor setting message given the excesses feels appropriate to like reframe the entire conversation because until we do that, I think the chances of actually emerging with the like balanced, meaningful value laden. Path you're describing seems like it just will never happen. I don't think we can incrementally get to it being good, but maybe the impending economic collapse and generative AI bubble popping will do some of that work for [00:28:00] us. I don't know. Um, but I hear you that like at least we should be able to articulate that vision. 'cause if we don't, then it will be co-opted by someone who pretends to have that vision, which I think is probably already happening. Chris: I was just gonna build on that, but also agree and agree with both of you now that yes, there is a positive vision and I think it's good to speak that into the world so that we can like continue connecting both sides of the right and the left and and the center and whatever. But I do think it's important to then focus on limiting. The excess of data center building and AI in particular, building environmental policy that is actually focused on data centers and doing that by changing the legal definition of data centers and actually naming them as a polluting industry. That is part of what people are speaking about now, how data centers are becoming the next oil and [00:29:00] gas industry. Prathm: I think that there is. Consistent in a lot of the case studies, these differences in what the explicit harms are to a community and they stick out the most when they are part of a legacy of similar harms caused by different organizations and people. In the South Africa example, this is a. Country with a history of drought, a history of water shortages, a history of power outages, and then to build a data center in communities that are already facing, that is taking that to the next level. And the Chile example of doing it in a country that is in experiencing nationwide drought for so long to build a water hungry data center. It's so obviously the continuation of harms in a way that in other contexts, when we think about a data center being built and people's main concern with it is it's, you know. Ruining of the natural beauty in that area. While that might be a valid concern and for a lot of reasons, I think that the difference in harms and you know, who faces the brunt of the most serious ones, is a thing that is repeated in the data center [00:30:00] struggle as it has been in plenty of colonial struggles before in the past. Chris: Exactly. And I think that's also where we're seeing. Data centers being built the most is they're actually being built in drought ridden areas. And that's the question that I always have with data center companies is why are you building it there? There's like absolutely no consideration for the environment or the communities, and I think that needs to be built into a consultation. With communities when you're building a data center. If we could dream up ways to change policy, but I think that's a really big factor, is at least having an environmental impact assessment, which they did do in Volve. The ultimate undoing was the government decided because meta couldn't perform on any of the environmental obligations, they could not build a data center there. Alix: Yeah. I also found one of the interesting things within the Carrero case study was that acting like data center resistance is [00:31:00] a totally new group of people that are fighting something totally new is just a historical and not particularly useful. And that actually, if you are. Coming at this from a sort of technology politics angle, you might miss the, like, decades of organizing around water rights, around indigenous rights and indigenous sovereignty, or against oppressive regimes and authoritarianism. And I feel like there's a sin that gets committed a lot in this space of like columbusing, uh, problem and then being like, who's gonna work on it? And it's like, oh, there's so many. Communities, so many harms, so many sort of accumulated political histories that we have to understand to be able to understand how best to think about what should happen next in terms of resistance. That case study was the clearest in terms of their being these like intertwined communities and organizers and people fighting different things to varying degrees of success. And also like there's a presumption in the Netherlands example [00:32:00] that like the alderman. Should have appropriately disclosed what they knew about what was happening. And like that there should have been good governance because it's the Netherlands, which is also this kind of presumption of good governance, which doesn't exist in the Netherlands either. But in Mexico, obviously it's a lot starker in terms of, it is naive to presume that a state is acting in the best interests of the people in a community and being like. All we need is a checklist for local policy makers to better understand like there's a presumption of innocence. I think in terms of political contexts that we shouldn't make when we have these conversations. Local policy makers sometimes aren't on the side of communities and we need to think, be real politic, um, about that. Prathm: I don't think we should discount. How much of this could be an information problem? I take the point, obviously, you know my politics, there are policy makers who do not. Have the best interests of [00:33:00] communities in mind, but I think that there are also policy makers who do. I think that there are mayors of towns that have invited data centers into those towns out of ignorance, not out of some sort of deep sense that the economic conditions of. Bringing companies in is more important than the lives of the residents. Obviously there are people who do this too, but I think there are a lot of context in which it is the case that all forms of resistance are necessary in order to prevent harm like this. But I think that there are examples, and I think that we've seen examples of like just simple political pushback. Informing politicians getting city councils to redo their votes have been helpful after the context this gained. And we see this in political issues of all kinds, but this is a complex topic for a lot of people. A lot of people have not thought about the potential environmental or economic outcomes. A lot of people have been misled by claims that data centers actually create jobs when we know they. Barely do and do not trust anyone who's telling you like hundreds or thousands of jobs are gonna be created by a data center, maybe in construction briefly, but there's not a lot of work that [00:34:00] goes into maintaining a data center. And so I think that politicians, policymakers, in some cases have just been misled by this information. And it's actually just like a counter lobbying problem. In some of those cases, that would be a result of more information transparency. Alix: I think that's true. I think it's an information problem, but also a narrative problem. Sure. Um, and I've been thinking a lot about like. The hangover of thinking about tech growth as green growth, this feeling that like expanding your economy based on tech is an infinite positive. Good. Because it was thought of as kind of like one, the like political engagement of tech companies in the early aughts. The left, I think associated in their head that those companies were like. Good companies and they haven't quite evolved that thinking to be like, oh, Google's just like General Electric, or Proctor and Gamble now. Or like Lockheed Martin, like, we should be treating these companies like just normal companies that are profit motivated, which means they do all kinds of messed up stuff because they're pursuing profit and they need to be held accountable and contained, and government should be skeptical and responsible and kind of negotiating relationships with them. That's the first thing is I think that there's like this hangover from like the politics of these [00:35:00] companies. And the second is that. It's oftentimes been framed as like a services economy. So it's like transacting with Amazon increases the sales tax in my state because when someone buys something online and they send it here, we're getting sales tax, for example, and it doesn't cost us anything and we get revenue from these companies. And I think there's like a kind of naive understanding of governments. Before the like massive digitization process that has required all this physical infrastructure where they're like, oh, if we could convince a tech company to come here that's like free tax revenue, which also kind of then starts getting into an information problem where it's like they don't understand the nuts and bolts of these deals and also are probably like. Feeling inadequate in some way. Like, oh, the tech company's gonna come in and they'll tell us what the deal is and we'll just say yes because it looks good to consumers. It looks future oriented in a way that I think is also attractive. So I feel like there's an information like nerdy policy set of information that's missing, and then there's the like bigger narrative of like how policy makers feel connected to some bigger vision of like what they [00:36:00] want for their communities. And I think that both of those could be good faith. Animated and fixed. Chris: There's also like an education issue, and I think that data centers need to be put into the mainstream more, and especially in science related to climate change related to the environment, related to if you're taking anything relating to computing or technology studies, even business, it's like data centers need to be the next climate change in terms of like how it's. Incorporated into every part of the curriculum so that people understand what it is and where it's coming from. Also, for example, in ero, no one in the surrounding communities of the data center knew what a data center was. They didn't know what AI was. They didn't care 'cause it didn't serve them. So I think there's also this issue of. Maybe in the US or in places where people are using technology more or they're interacting with AI more, it's more sustainable to [00:37:00] educate or speak with the communities. But in places where it seems like it's more of like an extractive. Extracting the resources of a community without even interacting with them. I think there has to be like alternate ways of figuring that out, interrogating that situation. Prathm: Just to highlight that point, there's like plenty of philosophers and ethicists and activists, critical scholars who have written about this idea. But the thing about data centers and the cloud as a concept is the kind of inherent abstraction away from. The actual technology, you know, the computers are elsewhere and you access them through the magical internet, and so you don't really think about the physical location of these things and the impacts that they may have until they're being built in front of you or taking your resources away. And even then you might not know that it's happening. And so there's colonial legacies of abstracting the harms away, such that people do not think through. Or don't have the opportunity to think about where the harms are actually happening because they're just a little bit further away from them, and that is a huge advantage for the people. [00:38:00] Building harmful data centers is apart from the education problem, the education problem is not just about informing people about the data centers, but also having to get them to feel their actual existence, like have them as a part of their phenomenological experience. It's not easy, it's delete that. I can't use the word phenomenological. Alix: Academic, uh, side is like a Jack in the box. Yeah. Um, where you're like, da phenomenological Prathm: that, delete that is because I was reading neuroscience, delete. Um, Chris: no, I, I thought it was, I was just like, whoa. That like shook me to my core where I have to. Alix: Figure that word out, that's a academic research methodological lens on the case study structure. It's like grounded theory. So it's like not making presumptions about what's happening in contexts and starting with who is doing what and how can we learn about what happened. And then from there, trying to synthesize what we've learned and like make presumptions. And I feel like. Generally speaking, like none of the conclusions we came to were that surprising. But I think because we started where we started, I'm much more confident that [00:39:00] they're correct. And I feel like that's like a grounded theory thing. And then phenomenologically like. Recognizing, trying to boil this all down into something is inherently reductive and actually every community experiences this differently, which is part of the strategic challenge of figuring out what are the collective actions that have high leverage that could help us get ahead of the corporate global approach. That basically is hard to combat when the companies are acting as one across lots of contexts. But it's impossible to imagine or difficult to imagine policymakers. And communities being able to compete with that, both from a resource standpoint, but also just from like a knowledge. 'cause they're like each data center they build, they're learning more about the tactics and the process and the costs. And because they're not required to share any of that, they're just at this like increasing competitive advantage in relation to a phenomenological experience of a local policymaker who one day someone knocks on the door and is like, Hey, we wanna do. Something here, and it's hard to know that when those moments are happening and [00:40:00] to know what can be done in anticipation of them that isn't centralized or something. Chris: Yeah, and I think in part of, or like a reason why, one of our recommendations was to increase convening was to also share knowledge. And when we're speaking about the policy maker that might not fully understand why they're lobbying for an increase in data centers in an area. If we're bringing them in conversations with community organizers that know exactly why a data center should not be in this area, maybe they're gonna also get informed. Maybe they can share knowledge with how policy is actually being created. And I. I mean, maybe this is too big rainbow, everyone coming together, but I really think that there are ways to strategically place people so that everyone feels like they're able to move forward with their needs being met. Prathm: It feels like a baseline recommendation, right? Like people just have to have access to information before any of this. Knowledge, sharing information, et cetera, at, at no point are we saying we solve this problem if everyone knows what's [00:41:00] going on, but like, yeah, at the bare minimum, we need to know what's going on and we need people to be able to communicate with each other. And it feels like we're so far away from that as our reality right now that we have to start there. But there's so much more to do after. Alix: I mean, all this requires resources and the only people that put money into community organizing and resistance that isn't companies is private philanthropy. In the report, we have recommendations for private philanthropy in terms of where they might grant resources. Do you wanna say a little bit about what we recommended for funders to be thinking about? Chris: One of the recommendations was ensure that community consultation is meaningful. And I think that's also really hard to define or parse out. But there also has to be some like careful consideration of how to interact with community organizations if you're a policymaker, because a lot of the time policymakers are not entirely trusted, or funders are not entirely trusted. So I think also providing convenience spaces that feel like, like you said, Alex, like the community is centered in the [00:42:00] approach so that you can have these like meaningful conversations. That's a, a way to move forward where you're kind of breaking down the silos of everyone's individual expertise to share expertise horizontally and have like a lateral hierarchy. So one of our recommendations was how funders could provide material support for communities resisting data centers. This would be in terms of funding or in terms of time and how they could listen to community needs and how they can support their advocacy work. And this is part of the ensuring community consultation is meaningful. The best advice I've gotten from working with communities is from a mentor, Dr. Warworth, and he told me that. You have to do what's best for the community, for what they're asking for, and not just do what you think they need. And so I think that is part of how you can interact with communities in a more meaningful [00:43:00] way. Prathm: The funding recommendations can be boiled down pretty simply to like, I. Ask communities what do they need in help and in money and resources, obviously that is the primary one. And then some of the things that we have found in our own work that might be helpful alongside, or you might hear from communities themselves, are funding things like strategic litigation, funding transparency efforts, especially things like investigative journalism. When we had investigative journalism, I think that was helpful for many of the communities and also helpful for researchers looking at it. And then, yeah, funding the convenings and, and helping people share information together and work together to. Solve these issues. So maybe the summary that I want to give here is. I'm sure that many listeners are realizing now that the story we are telling is a very common story in a lot of ways, which is giant mega corporations are doing some sort of harm that they promise is for good. That harm is mostly located in the most vulnerable communities who've already been exposed to harm. In order to resolve this issue, we need to support the communities with actual [00:44:00] resources and with time and with attention, and with information sharing. And we need to fight the structures of our political systems that allow for these harms to continue and often incentivize these harms against the most marginalized communities. All of those things that we have seen time and time again in the eras of capitalism we are seeing with data centers. It's just that some of these harms are really stark, and so the recommendations for how to solve those are also quite similar, which support the communities with money, resources, knowledge, information, and time. To Chris: build off of that, we also had a few recommendations for the data center companies and government leaders in general. We wanted to make sure that data center companies know that we know that they're holding their information to. From us, and so we're asking for them to make key information available for everyone and have them distribute information about water use, energy use, and their environmental impact assessments and make it [00:45:00] publicly available. We also wanted to understand what information government leaders need to make these informed choices about if they want to build a data center. What is the environmental impact? What is the community impact? And we want to include this sort of training or education for government leaders to be part of our recommendations as well. Alix: So I feel like maybe a good place to end is thinking if in let's say five years this was all happening in the way that we wanted it to be happening, what would it look like? Prathm: Uh, I, I want to answer this directly, but I'm going to do the same thing we just did, which is I, I don't think we can decide that because I think the communities in which these data centers are being built in have to decide. If they want them, what they want them to look like and how they want them to impact or improve or modify their lives. So in an ideal scenario, data centers, if they are still being built, are being built [00:46:00] in mind with the people who live there, who have real power and control over the situation. Whether that means, you know, running it themselves through cooperatives or publicly owned ones, or through public-private partnerships of the kind. Chris: I heard a cool thing, which was like, in five years there's gonna be an AI crash and everything will be obsolete that we have built in the last 10 years, which would be an interesting way to look at. Future. 'cause then we are gonna have these boxes everywhere. So it's like, are they gonna become homes? Are they gonna become other businesses? Like what's gonna happen with that? That's a future that I could think of, but in a more realistic setting, I agree that I am hopeful of the positivity that Ham is talking about of communities being able to directly consult with governments or data center companies having. Data centers be less of a intense topic. It's kind of funny how now they're like, it's on everyone's lips [00:47:00] data center, where before this wasn't a topic that wasn't mainstream, it wasn't in the news. But now we're seeing that Trump is investing like $500 billion to Stargate, and we're seeing that the UK is now making data centers like their critical digital infrastructure projects. So my dream would be in five years that data centers don't exist anymore in the same way that we're seeing them now, and that they're used more in the, in the ways that we use them in the previous decades. But if they are being built, it would be great if they were smaller, if they weren't harming the environment at all. I know that there's some prototypes happening in Sweden where they're submerging smaller data centers into the ocean to cool them down. That also sounds harmful to the ocean. So I feel like at any point there is nothing that you could really do where you're constructing something on the environment where it won't harm the environment. So I guess I'm going back to my data center abolitionist [00:48:00] stance and I'm saying no data centers 2030. Alix: Yeah, I mean, I think between now and 2030 also the obvious. Manifestations of the climate crisis are gonna be such that it changes the conversation a lot. I mean, I do worry that the Eric Schmitz of the world will continue to make the case that digitization is a way of continuing to have growing economic activity without the emission costs of other types of industries. And also that it's like the only way out is through, and that the path of digitization and increased computational power is the only way there we're gonna. Find solutions to the climate crisis, which I think, I presume the three of us are on the same page that like that is a garbage argument. But I think it's a really resonant one. 'cause it's a very attractive proposition that there's a way out, even if maybe there is not. Chris: I think it's crazy that Sam Altman. Said that he wants to engineer the Earth to solve climate change. Actually, Alex, I remember one of the podcasts that [00:49:00] you had, you had on Concrete specialist who talked about how they were using AI to control the temperature so it was exact so that they could melt. Gravel and create concrete. And you were like, I think this is the only just way to use ai. And I totally agree because I think there are some energy intensive engineering work that goes on with mining and with fracking, whatever environmental harming stuff they do. But there are also ways to. Finite and control it so that it is not quite as energy intensive, but like those are the, but it's stuff that Alix: humanity relies on. Like I didn't realize for that. Like concrete is something that we so need, like if for our built environment, it's just, and so in this situation where we have decided as a human species that we need a thing, making that reproduction of that thing as efficient as possible feels like a really good. Thing to do. Prathm: I think AI is gonna help us design better batteries and maybe more [00:50:00] efficient solar panels and other materials that might be useful in our pursuits of helping the climate. But it's like, can someone come up with any other situation where it's like, this thing is causing a problem, we should make it cause the problem so much until it solves the problem. Like that's not how this works. Like obviously we might get some help, but AI's not solving climate change. Alix: Okay. I hope that was interesting to you. If you're working on data centers, we have an upcoming community call to talk more, to platform each other's work and just to connect across regions, continents, et cetera. And also sectors of people that are just interested in better understanding what's going on, what's the scale, what can we do to equip communities so that they can, in a more discerning way, make decisions. About whether they want these things in their backyard. I'm coming at you live from Memphis, where my family is, and also the home of XI, which is [00:51:00] a behemoth data center and supercomputer cluster that has kind of taken Memphis by Storm. Will link in the show notes to tech Policy Press' recent episode, interviewing organizers who are. Pretty upset about the lack of oversight and also the corners that are being cut with the building of that facility, which I don't think it's an accident, that it is run by a Musk company. Yeah. So these data centers are not mythical. I. Physical infrastructure being built somewhere else, you're being built in your backyard. In my family's backyard, it's becoming a more and more pressing issue for us to get a handle on because it is not happening with the oversight and the community consultation. There's a lot of rose colored glasses I think. Talking to family here. There's just deep frustration that local politicians think this is some shortcut towards prosperity, towards political winds that make it look like they're rejuvenating economies and creating jobs when that's just not the [00:52:00] case. They're essentially giving the green light to people like Musk to build environmentally devastating infrastructure in folks' backyards. And those people have had no say in whether they want it, and a lot of them have actually been subjected to many industries that have come before who have built infrastructure that has caused huge health problems, huge community pollution problems. When I'm at home normally. We don't talk about politics. Um, and it's been really interesting both to hear how top of mind the XAI facility is both in terms of a feeling that maybe we're sharing. I had an interesting conversation with my sister-in-law who said when she heard that it was first coming, she was really excited 'cause she thought it was maybe like the next wave of companies like FedEx, who. Created a ton of jobs who had a very like innovation forward look, and now she feels a bit silly slash betrayed slash It seems obvious that this was always a terrible idea. Now that she really understands the implications and the fact that it's not creating very many jobs, and it's basically just polluting. Already horribly polluted neighborhoods in [00:53:00] the city and without any consultation. So if you're wondering if everyday people who don't follow AI politics are following this, they are. And they're really pissed. Um, and I think that there is a lot to be done around this, both to connect communities who are affected by these. Projects. Um, and also generally make the work of AI politics more accessible so that more people who are directly affected can participate because it is now hit the mainstream in terms of people's understanding, as far as I can tell. And, uh, I think that really matters. So I hope this, uh, report was really interesting to you and that you, if you're working on data centers or interested in working on data centers, that you'll join us for our next community call. And just be in touch about your work. We will be continuing this work. A link to the report is in the show notes. Please share it with friends and organizations you think would find it useful. We hope it's useful to you. So thanks, and thanks to Georgia Iacovou and Sarah Myles as ever for producing the episode. Um, and we will see you next [00:54:00] time.