A month ago, Google released its 2024 sustainability report. Its overall water consumption increased by 28% in a year. Less publicized than the data center energy boom, water is also pivotal for the Tech industry and data is even scarcer.

To better understand this secret but serious love affair between big tech and water, what better location to investigate than its birth place? Yes. THE silicon valley. And no one there is better qualified to explore the topic than Masheika Allgood who lives in Santa Clara and recently created a Data Center Water Consumption Calculator based on public data. Over their conversation, Gael Duez and her covered:

The astonishing amount of water used by data centers in California and why it could have been even worse
“Back to the loop” & the limits on the efficiency gain from new cooling technics
Air pollution & the trade-off of heat reuse in urban area
The unignorable noise pollution
How big tech lawyers have an edge on city council and community activism

And much much more!

❤️ Subscribe, follow, like, ... stay connected the way you want to never miss an episode, twice a month, on Tuesday!

📧 Once a month, you get carefully curated news on digital sustainability packed with exclusive Green IO contents, subscribe to the Green IO newsletter here.

📣 Green IO next Conference is in London on Septebmer 23rd and 24th. Every Green IO listener can get a free ticket using the voucher GREENIOVIP. A small gift for your huge support. 🎁

Learn more about our guest and connect:

📧 You can also send us an email at contact@greenio.tech to share your feedback and suggest future guests or topics.

Masheika's sources and other references mentioned in this episode:

Transcript (auto-generated)

Masheika Allgood (00:02)

Okay, so let me say that a 100 megawatt data center, system is gonna require 223, I hate to say it, because it sounds insane, but 223 million liters of water in the system just to get the cooling requirements. That's the spin up the system. And then it runs through three million liters of consumed water every day.

Gaël Duez (00:32)

Hello everyone, welcome to Green IO. I'm Gaël Duez and in this podcast, we empower responsible technologists to build a greener digital world, one byte at a time. Twice a month on a Tuesday, guests from across the globe share insights, tools and alternative approaches enabling people within the tech sector and beyond to boost digital sustainability. A month ago, Google released its 2024 Sustainability Report. Its overall water consumption increased by 28 % year on year. The good news is that location-wise water withdrawal and consumption data were made available. Water being even more difficult to transport than electricity, how aquifers are used by its different consumers impacts a lot the local communities. And data centers are not the last to consume water.

Ask our British friends, who recently discovered, thanks to a combined work from Foxglove and The Times, that 231 English data centres consume the equivalent per year of almost 4,000 Olympic swimming pools. But they also discovered that this figure is completely underestimated, because over half of British water companies have no clue how many data centres they supply, nor how much water they are hoovering up. I'm quoting here Donald Comble of Foxglove. To better understand this secret but serious love affair between Big Tech and Water, what better location to investigate than its birth place? Yes, the Silicon Valley. And no one there is better qualified to explore the topic than Masheika Allgood. Masheika has already lived several lives. As a lawyer, as a techie. For instance, her team at NVIDIA wrote the software that enables AI training, (she built the enterprise licensing system required for running AI on your GPUs), and now as an activist. She recently made a big splash, pun intended, by creating a data center water consumption calculator based on public data and open hypothesis. So, welcome to the show, Masheika It's really an honor to have you with us today.

Masheika Allgood (02:50)

I really appreciate you having me on. This is a really cool opportunity, so thank you.

Gael Duez (02:57)

Well, I think we have, as I said in the introduction, I mean, you've lived several lives. I think I've got like half a hundred questions before once. And that might shock my usual listeners. I'd like to start right from the start with a provocative question playing a bit the devil advocate here. And no, I'm not going to turn into some clashes, stupid YouTuber, you know, to get more clicks and likes and what nots. But I would like to put this question on the table first, maybe to evacuate it fast. So you're someone who's deeply engaged. Again, the water consumption of data centers in the Bay Area, and we're going to develop this. But my question would be, do you consider yourself as a NIMBY, not in my backyard person, or actually that will be a very wrong way to describe what you're actually doing?

Masheika Allgood (04:00)

So I don't think this should be in anyone's backyard. So I don't think it's gonna be so separate. Yeah, so I live in California. I live in the Bay Area, but I'm from Florida. And Florida is one of the wettest states in the US. So I grew up with rain as a regular feature in my life. And I moved to California at the tail end of 2014, which was in the midst of like a 10 year drought cycle.

Gael Duez (04:04)

I was expecting this answer!

Masheika Allgood (04:30)

So my first couple years here, it never rained, like at all. It was freaky. Like I had friends who were like recording the rain and like sending me clips and it was like, my God. Like I just, I missed rain. had never been anywhere that just didn't rain. And it just, it was a surreal experience. And when it started raining, I was like, ⁓ this is what that is. Wow, how cool, right? And so I didn't understand the concept of water scarcity because I came from a place where there's ocean there, there's lakes and the intercoastals, like a river on the inside. There's water everywhere. When you drive, you always see water. And I come to California and like there was no drops of water anywhere. You're always hearing drought. Everything was brown. My wife laughs because I thought they called it the Golden State because it was brown. Because when we drove everywhere, like all the grass was like golden. I thought that was it. I knew the real reason, but for some reason my brain supplanted it and she laughs to this day. But water scarcity is a real thing in the US and in California and specifically. We've had massive fires over the last couple of years, all up and down the West Coast. These horrific like looks like the world is ending kinds of landscapes from fires. So water is a real issue here. It's the states. And so people don't necessarily consider those things. But when I realized and when I learned that data centers take drinking water, it just, all of my alarm bells went off and it became a cause for me. You mentioned in your intro, you called me an activist.

Masheika Allgood (06:24)

It's a term that I'm not fully comfortable with because I'm more of an educator of activists, right? Like I'm not all that great with people like those who like me, like me. I lead amongst, you know, in certain situations, but there are people who they can mobilize communities and they're true activists and I try to assist them in what they do. But with this, this is the first time I was pushing, like, hey, activists, this is something you might want to do ⁓ because it's of critical importance, not just in my neighborhood, but water scarcity is an issue all across the globe.

Gael Duez (07:06)

Thanks a lot for the clarification about not being an activist yet being very active, would say. so if we try now and, and thanks for the clarification about the not in my backyard, meaning actually not in anyone's backyard. I was expecting this answer, but I'd rather have you clarifying it.

Masheika Allgood (07:11)

Yeah.

Gael Duez (07:28)

Now, how come that it is such a pressing issue in California? mean, do you have a lot of data centers? Obviously, this is a paradise for tech companies, so we could expect, but it's very abstract for lot of people. What are the order of magnitude? Maybe some people know about Virginia and, you know, data center alley near Ashburn but how bad is it? Regarding the concentration of computing powers and how bad is it regarding the water consumption? Could you provide us some meaningful numbers just to grasp something that is pretty blurry in everyone's mind starting with mine?

Masheika Allgood (08:09)

So I live in Sunnyvale. I can walk. ⁓ I don't know how you do it in metrics, but I can walk about 10 minutes from my front door and I'm in Santa Clara, right? Santa Clara is the seat of Silicon Valley. The county I'm in is the home of Silicon Valley. There are 55 data centers in Santa Clara.

Gael Duez (08:18)

hahahaha

Masheika Allgood (08:37)

Along this one little narrow strip, like this is the home of tech, which means data centers started here first, right? Because you needed them to build the tech companies, right? It's symbiotic relationship. The 55 data centers we have in Santa Clara makes us the largest concentration of data centers in the US outside of Loudon, Virginia, which is number one. So it's, we talk a lot about second in the US where we're tops in the world, like we're in the top 15 in the world in terms of concentration because the US has way more data centers than anyone else. So us being the second largest in the US makes us a player on the global stage when it comes to data center concentration. The difference with our Santa Clara data centers is we've been doing this here for decades, right? And so the vast majority of those data centers are not AI data centers. We're more recently building AI specific data centers, which are the ones that are most consumptive. Also, California and we do some things right here. So we have a cap in California. You can't build a data center larger than 99 megawatts. Companies are trying to get around that by building like two or three 99 megawatt data centers on a campus came. But we'll never build anything as big as say the massive one they built in Memphis ⁓ or some of the ones they built in Atlanta. The big super huge mega data centers cannot be built in California. Our structure doesn't allow it.

Gael Duez (10:14)

How come that this cap exists because it sounds a bit counter-intuitive in the country of big tech? I mean, this is not the first time I hear it. You have it in Singapore until recently. It was lifted recently, but you had it for several years. I know that in Ireland, the electricity regulators basically put a ban on any new data centers if they're not self-powered which brings a lot of questions regarding building gas powered how come that this legislation was in place for less than 100 megawatts?

Masheika Allgood (10:51)

So California is an anomaly and I think the world should be a little grateful that the tech started We're anomalous in the US when you look across our spectrum of how we handle industry and politics. So California has to a greatest extent tried to walk the fine line between innovation and responsibility. We aren't the best at it. We learned some things during fire season that we've not done all that great with our water. But industry-wise, we have tried to be reasonable and walk the line because California is one of the bread baskets of the U.S. We grow more food here than almost any other state. We have some of the most fertile land in the U.S. And so you have to make provisions for farmers to have water as well as, you know, industry.

Gael Duez (11:42)

So the cap was related already to water consumption more than energy consumption, the 99 megawatt.

Masheika Allgood (11:52)

I might have say that because we have a variety of industries. So it wasn't just a water thing. That's just my focus of it. That was the benefit. ⁓ But no, we've we've had the cap for a while. I wasn't part of that legislation. I can't speak to the impetus, but I know that we in a variety of areas have been trying to walk this line between innovation and responsibility to the public and to other industries. So because we're

Gael Duez (11:54)

Okay. Okay, got it.

Masheika Allgood (12:20)

We're a massive state. We've got a lot of competing interests. We've been pretty good about not allowing any one interest to take all the things. So our farmers are always protesting they don't have enough water, right? Our techies are always protesting they don't have enough power. Everyone's always protesting, but everyone's always getting business done. So like, if everyone's unhappy, you're probably doing it right kind of thing. So that's kind of where we landed. I don't know why they chose 99 megawatts, but I am so grateful that they did. Anything larger, we'd in a much different situation.

Gael Duez (12:53)

Let's go back to the water consumption ⁓ 99 max megawatt data centers, do they consume a lot of water? this is a question that was actually asked by some listeners on the Green IO Slack workspace what is the cooling techniques and how do you explain such a big water consumption from data centers? Specifically here in California or in Santa Clara.

Masheika Allgood (13:19)

Okay, so let me say that a 100 megawatt data center, oh, let me put it in liters for your audience. system is gonna require 223, what's that? Oh, no, that's pretty crazy. I hate to say it, because it sounds insane, but 223 million liters of water in the system just to get the cooling requirements. That's the spin up the system. And then it runs through three million liters of consumed water every day. So like the system cycles through that 200 million and then as it cycles through it consumes either through drift, so the little particles flowing away, evaporation or blow down, which is like the water that has gotten too heavy with sediments or too polluted and they have to like push it out of bottom of the system. So that's like 1.5 % of the water that is in the system every day is consumed, right? And so every day it's consuming around three million gallons of water, or liters of water, I'm sorry, gallons is different. So I got those numbers from my data center water consumption calculator that I built. I basically, no one had any numbers and I was like, someone has to have numbers because you can't keep relying on the tech companies because they're not being forthcoming. So is there a calculation that isn't specific to that industry or isn't specific to that particular data center? Then you'd have to like talk to the operator. But is it like some wild fantasy numbers? Is there something grounded that is respected in the industry? And the Uptown Institute is an organization that basically certifies data center sustainability. And it's not sustainability as in you're good for the planet. It's in sustainability of operations. Can you operate in case of loss of connection either to the grid or the public water supply? How long can you operate? So they have this makeup water calculation, which basically determines if you can keep running within 24 hours, right? or for 24 hours or for 12 hours, how long your systems can run without connection to water. So it's at the design phase and it's based on the power output of the system. So there's some assumptions made, right? And we can argue those assumptions, which is perfect. Like that's the point. You want to be able to argue those assumptions and tell me that your stuff is more efficient. But the assumptions are reasonable assumptions that are accepted within the industry. And based on those reasonable assumptions, that's what the water calculator calculates. So those are your baseline numbers. Your question was, what about more efficient methods? And that's where things get interesting. ⁓ So there's really kind of three different methods that people are touting as efficient. And your first is going to be direct liquid cooling. So instead of… running a closed loop of liquid through the back of the server, I actually put a little connection on the back of the GPU and I squirt water specifically on the hot spots to try to cool the hole because some spots get way hotter than other spots. So I just shoot water directly on it. So it's less water intensive. So that's the first option. The thing about that option, so I'll tell what the option is and why we have difficulty determining how efficient it is. It's supposed to be significantly more efficient, right? But we don't have numbers at scale, right? Like we're not looking at 100,000 GPU installation, which is what some of these data centers have. So we don't have numbers at that level of scale. ⁓ And I'm not exactly sure like if that water is reusable. So they're shooting the water in, it's going into some sort of repository and then being cooled and then coming back to the system. But if it goes through that kind of process, then it's going to have the same problem as any other liquid cooling, which is there's a heat exchange loop. So the loop where you're pushing up against the server or against the chip is closed. That is a closed loop, nothing gets lost in that process. But you got to cool it. At some point, that water, that liquid goes through and gets cooled. It's that cooling loop that's the problem because typically that's evaporative cooling. We all know evaporative cooling is massively consumptive and that's essentially what the calculator is built on, is evaporative cooling. If you have a water tower and use a chiller, then this is what your water consumption will be.

Gael Duez (17:53)

Okay.

Masheika Allgood (18:07)

If your new efficient system still uses a water tower, it's only so efficient. And so we're not getting the numbers of how, what's the efficiency of the cooling loop, right? We're just getting, but it's using less water on the chip. So that's one new method that we just don't have numbers on.

Gael Duez (18:28)

I'd like to pause here and to rewind a bit just to make things clear for everyone. So in the Bay Area, obviously, you don't use air cooling. You need water cooling, I guess. And you mentioned it several times, but most of the time, this is closed loop cooling, which means that the water is getting reused several times. And then at some point when it's getting either too hot or too polluted, it gets throw back into the system. Is it correct? Just to clarify.

Masheika Allgood (18:59)

Okay, so first off, we can't use air cooling in most of the world because you need the temperatures outside to be cooler than the temperatures in the data center. And I don't know if anyone's noticed, it's really hot in the world right now. So that's tough. So that's like free air cooling, And then when it comes to actually like, I'm running fans, it's very difficult giving the heat these GPUs put off.

Gael Duez (19:04)

Yeah. Really?

Masheika Allgood (19:23)

for fans to be very efficient. So it becomes very electricity intensive. So there's this trade off between electricity and water. I use less water, which means I use more electricity and we know carbon and that has its own issues. this is why I say, ⁓ always, always. So there's no math reason for them to make that choice, but then it doesn't benefit society if they do.

Gael Duez (19:25)

Plus cost, guess energy costs are higher than water costs for many operators. Yeah.

Masheika Allgood (19:50)

So this is why I'm not in anyone's backyard. There's no good answer for this, right? So, okay, so it'll go through the loop and that loop is closed. The issue is you have to cool the loop and you can either cool that loop through air, which is more electricity, or you can cool it through water, which is water consumption. So in California and the Bay Area and pretty much in every data center, you have this issue of how do you cool the loop? And you really have two options, kind of three if you live in a colder environment, let the air cool it. But very few data centers are in that area. And so that whole, when the water goes bad, it goes down, that's through the cooling loop, right? Because the other loop is closed. So we're not throwing any of that away until it's like ridiculous, right? But when I'm evaporative cooling, then that system has a loop, right? And we're pushing water through and it goes to the top and it evaporates and we're just forwarding water on the other liquid to make it cool. It's a whole process. But yeah, that loop, some of that water gets recycled and you have to Blow down that water. there aren't very good answers, right? So if I use direct to liquid cooling, that's less water in that closed loop. But, the questions are, so how does that recycled? Like, is that water cooled? Like, we just don't know enough about the entire life cycle of that process. There's also the, can I put it in the ocean or a deep body of water? That was tried by several different companies, but the chips run so hot they were impacting the ecosystem in the body of water. So that can't be, you can't be heating up, you know, your cool lakes and killing all the fish and birds. So that didn't work out. So then it became, can we cool it in a body of liquid? And that is kind of like the new thing. I submerge it in a body of liquid? And so there's, yeah. So there's, there's issues of PFAS chemicals, so like the types of chemicals they submerge it in. So your environmental folks are worried because if that gets out into the environment, it's disastrous. But there's also the issue of you can submerge it, but it gets hot and it has to be cooled. So you still go through the process if you have to have a cooling system. when we talk about more efficient, I don't doubt that they are more efficient than just straight evaporative cooling but we don't have numbers on what more efficient means. And this is the danger, because people say more efficient as if it's efficient enough. But since they're not telling us how much water they're pulling from specific aquifers, we don't know what efficient enough is, right? So like if you're 5 % more efficient, but you're overdrawn the aquifer by 20 % of its capacity, is that sufficient? No. But can we do that math externally? No. And so this is my concern with efficiency methods, is they're not based on anything.

Gael Duez (23:01)

And there is a clearly a full line of questions that I'm going to ask you about having access to this number of transparency and so on. But I'd like to stay with this efficiency discussion for a bit longer, because that was also a question that was pushed by one of my European listener in Europe.

Masheika Allgood (23:15)

Thank you.

Gael Duez (23:21)

and especially in the coldest place of Europe, we tend to see heat not necessarily as something to be wasted because heat is precious and there are quite a lot of urban areas that could benefit from this heat. And it's not rare now, it's not the majority, but for many new and not necessarily big hyperscaler but like medium sized data centers that they are coupled with a urban heating or, you know, in Paris, they've got a solution heating a swimming pool. I that in Geneva, there is a full district being heated by one data center. There are several examples in the Nordics as well. and this is actually something that you can see in the European energy efficiency directive that how much heat is reused is now compulsory to be reported by data centers because this is something precious. This is energy that we don't necessarily have to waste in the atmosphere. Now, that being said, I live on a tropical island. You live in a hot place as well. This is not necessarily something that we need, but is there any way to be more efficient by not wasted this heat and maybe use it for other human activities like, I don't know, the industry, agriculture, whatever, or is it something that is absolutely not considered?

Masheika Allgood (24:16)

No, we're considering it, but there are problems. So San Jose is the city below me. So we consider it like a tri-city area. Our county is three cities. So it's San Jose, Santa Clara, Sunnyvale. There are some other cities, but for the purposes of this discussion, those are the three big ones. So San Jose is looking to build housing slash data center complex in the middle of downtown. The issue that we have with it, I am advocating very strongly against it because they were like, but we can reuse the heat within the housing for a variety of purposes. We can put up a greenhouse. It all sounds great until you recognize that all of these data centers run generators, gas or diesel generators on site. And those generators are in California only used for backup power. In Memphis, they're powering the data center itself, right? And the difficulty with generators is they expel small particle pollution, fine particle pollution. So the Bay Area, our Tri-County, plus San Francisco and Oakland, when it comes to air districts, because we have a very unique kind of way that air flows in the South Bay or in the Bay Area, ⁓ our American Lung Association, it's a national organization within the US, they keep a list of like areas that have really bad fine particle pollution. The Bay Area is number seven on that list in the nation, in the entire US for small or for fine particle pollution. So you put up a data center, it has generators, right? The generators are allowed to run, for maintenance purposes.

Gael Duez (26:13)

Okay.

Masheika Allgood (26:29)

They have to run a certain amount of hours every month, right? So think of 299 of generators.

Gael Duez (26:31)

They have to.

Masheika Allgood (26:41)

Running monthly in a city or an area that's in a valley, we get these, I forget what they call it, it's like air inversions where basically the air kind of sits over our valley and nothing gets in or out. So the pollution just sits, right? I was talking to someone who had covered the fires and they were like, yeah, so it sits. And then we always knew when the fire would explode, because the air inversion lifted, right? So when you're talking about reusing heat, if you're in a country that doesn't allow as much generators as backups, that might be a viable option. But in the US, generators are not only backing up data centers, they become first source power. So you've got a lot of data centers, which is what they're doing in Memphis. They're building… in areas where the grid isn't sufficient. They can't give them enough power for their operations. So they just use generators as their main source of power until four years later when the grid operator has enough megawatts for them. California, at least in the Bay Area, our data center operators have been building responsibly. Basically, they build as power becomes available. But that's not the case across the entire

Gael Duez (27:39)

That's insane.

Masheika Allgood (27:58)

There is a run on generators right now. Like electric companies can't get generators for backup because the tech companies have bought so many of them. So they are powering data centers all across the globe with basically dirty energy, right? And the environmental technology you put on top of a generator to make it less polluted, there's a run on that as well. So they're getting the generators without the environmental technology. So there's a lot of irresponsibility going on around how generators are being used. But even if you do it responsibly, we have 55 data centers in Santa Clara, right? and we've got another 15 or 20 in San Jose, and now you want to build 299 megawatt data centers. All of those have to cycle their generators once a month. So that's thousands of generator power cycling.

Gael Duez (28:38)

Yeah.

Masheika Allgood (28:55)

Every single month. I don't know how much heat makes that worthwhile. Like the air quality is staggering. It's insane to think of us putting that kind of load in a city that already is very high fine particle pollution. So yes, it makes sense when it comes to, we don't want to waste heat, but you have to consider, like none of this isn't a bubble, right? So you have to consider, well. What is powering it? Is air quality a concern? Because if you're having people live in a data center or right next to a data center, because it can't be too far, otherwise you'll lose heat in the pipes, right? So there's only so far away it could be for you to be able to capture that heat. Are there other concerns about having the data center located that close to people? It's not just about heat or electricity. There's a lot of concerns around data center.

Gael Duez (29:50)

That's almost a textbook example of pollution transfer and the necessity to have a systemic approach on the ecosystem surrounding a data center. So I guess, but let's not go into this direction, but we could argue that maybe there are better ways to capture the fumes and avoid this air pollution. But let's be honest, today, this is not the case. And while we are talking about

Masheika Allgood (29:55)

Yeah.

Gael Duez (30:17)

This pollution transfer, You told me when we were preparing the episode, another vibrant example of another pollution that is completely understated and you had a clear example, an anecdote that you told me about. Yeah, there is the issue with the energy consumption, the water consumption. mentioned the air pollution, but actually there is another player in the long list of pollutions and negative side effect of having a data center next to you. Can you bring us the anecdote?

Masheika Allgood (30:49)

Noise pollution is brutal for data centers and no one talks about it because it didn't used to be a thing. But data centers, the generators and the cooling, it all puts out noise and it's at a frequency that is at like human talking. So it's not ignorable. So a couple months ago, I was sleeping and I was awakened by this sound. It sounded like

Gael Duez (30:54)

Noise pollution. ⁓.

Masheika Allgood (31:18)

Electric vehicle, you know, they got the weird engine sound. So it sounded like an electric vehicle, but it wasn't moving. And it went on for like an hour and a it's three in the morning. And I had to leave and like try to sleep in the living room. It was brutal. like, you I went to a county meeting around data centers and the number one thing that the commissioners were talking about is can you handle the noise? We're getting complaints about noise. Can you handle the noise? Headphones don't work, earplugs don't work. It is pervasive and it's brutal. And there are things you can kind of do to help it, but a lot of that stuff sits on top of the data center. So like you can't wall it off. Like it's a, it's a really difficult issue that's becoming more and more known, the larger the data centers get and the closer they get to humanity.

Gael Duez (32:10)

And that's a good point that you raise that the logical solution might be, and actually for safety reason as well, let's ground, like, put these data sensors underground, but actually most of the noise comes not from the computing the processor, but more from the heat management. And obviously you need to be able to expel the heat, whether it's with air cooling or water cooling or whatever. And so the noise comes from all the fans, if I understand you well.

Masheika Allgood (32:37)

Yeah, the fans and the make just really frustrating noise.

Gael Duez (32:40)

The fans and the generator. ⁓

And Masheka, in a previous life, ⁓ you were a lawyer. Am I getting it right? yeah, you were a lawyer. ⁓ When we see all of these downsides and the fact that it seems to affect that many people, the logical question is how, especially in a free country like the US, where you've got a vibrant community scene How come that the data centers seem to have a free pass and they can basically pollute the air, emit a lot of noise, noise pollution, as you mentioned the water, seems that it's very cryptic. It's almost a secret. It requires you spy skill to know how much water is consumed. So what is going on with transparency and accountability from these big tech actors in the US or maybe just in California if you don't want to generalize?

Masheika Allgood (33:33)

Yeah. So lawyers are important, but right now the ones who got the head start are the tech lawyers. like ⁓ they understood what was going on way before anyone else did. They had lobbyists in place. And so that's why you see a lot of tech favorable bills that are getting passed in different states. They've been very active with NDAs.

Gael Duez (33:55)

Hahaha!

Masheika Allgood (34:12)

And telling leaders that they can't speak about things like water usage and consumption and electricity or when they're building and how big they're building. And when you're on the other side, your lawyers don't make as much. You don't have as many of them. Like you're asking for a city or county solicitor to take on a trillion dollar company. It's not that we don't do that. Obviously we do the underdog thing in the U.S. It's well known, but it's underdog for a reason, right? Like those are really hard fights to win. So you've got nonprofits and advocacy organizations kind of tackling that. And we're tallying up wins, but you know, there are little wins here and there, whereas there's a deluge of money coming from the other side that has been doing this for some time. And you can see it by how, know, Texas kind of ran over a couple countries in Europe and Africa. Well, not so much Africa yet, but that's next. ⁓ But they've come in and really sold themselves as being something other than what they are and really started trying to move the laws in their favor. So I mentioned Africa because they recently passed a law. ⁓ I think, I can't remember which country. I think it was Kenya. They have a law. Yeah.

Gael Duez (35:33)

I was about to mention Kenya because Kenya tries to a ⁓ lot of big investment in data they've got very low carbon electricity mix. So they try to attract a lot of investment here.

Masheika Allgood (35:43)

Yeah, and they've got a very highly educated workforce. Yeah, they're well suited, but they just recently passed a law that basically criminalizes letting out information about water usage of corporate actors. So we're seeing the playbook from the US play out in other countries. And the other issue is there are 300 million people in the U.S. We've got a lot of land. So a lot of stuff happens to groups of people that don't make it to the national stage. So there's this like simmering ground swell against data centers and you're starting to hear it pop up. But like it takes a lot for stuff to hit the national stage in the U.S. unless it's absolutely catastrophic. Right. And we haven't had an aquifer in a major city fail. As of yet. ⁓ And because AI has become an identity marker, like you have to overcome people's love and devotion to the tech in order to have real conversations about, you know, what it's doing to us while we're loving it, right? It's an abusive relationship, essentially. And it takes a lot to overcome, you know, people who are in love and abusive relationships. So that's kind of where we are. So you know, until something either catastrophic happens or something bad enough happens to enough people, it's very hard to get the communal groundswell. And let's just be honest, we got a lot of shit going on in the States right now that you can be up in arms against. It's hard to pick and choose your fight because everything is on fire, right? So real question in

Gael Duez (37:21)

I got Figure or that being said, you mentioned on the national level but as far as I understood, the US is a federal state and how come that seems not moving? At least in California where you mentioned it's slightly different, the vision of common good might be slightly different. The population experienced fires to an extent that was almost existential. The drought is something that you mentioned it right at beginning of the episode, something that is lived by everyone all the time. So how come that this topics doesn't come that hot, pun intended, in at least California? is it only explainable by the love big tech and how proud people are to be, you the fifth economic power in the thanks to all these big tech industries being

Masheika Allgood (38:20)

So first I'm gonna say, I think our main issues are, we haven't been able to tell the story, right? The fires were largely caused by our electric company, right? PG&E, they paid millions of dollars of damages to people because it was, you we have a direct actor, like they're power lines sparked and that's what caused the fire. So people have a place to put the blame for that. And so it doesn't come back to data centers as a story, right? We know that the US, when it comes to climate change and carbon, it's been very hard to tie that to individual actions every day. I think it's a marketing problem, right? Like we haven't been able to explain to people in a real way that this is a data center issue. Part of that is because we work for tech companies The Bay area, California works for tech companies. if you're not working for one, your cousin is, your friend is, you know, someone who is like most of the tech, jobs have been out here for a very long time. So there's this assumption based on the marketing of tech that we're the good guys, right? Like we've cultivated this marketing over time. And so no one wants to believe, it doesn't even occur to people to believe that what we're doing could be ruining all of us. Like it's such an opposite of what we have been taught and what we've come to know and believe about these companies. Mind you, the luster of that has started to fall off a bit.

Thank you, Elon, for showing the world who we really are in a real way out here. But there was this belief for a long time that tech giants are altruistic and these guys should be running these companies because they're just good human beings. And it's become clear that they can have problematic views. They can be bad people. They can do bad things. And we shouldn't just assume that because they run a tech company that they are good, but those were such coupled identities for so long, right? it's starting to become a thing where we're seeing them as but for a long time, they just rode the wave of tech is good, so they must be good, right? because we don't have real numbers on water consumption, we don't have a real understanding of how carbon for data centers impacts the heat and fires since we can't make a through line of that story, people just haven't seen. But you were asking for positives. I was just in a meeting yesterday with activists and different people who are trying to figure out how we tell that narrative because people are starting to understand it. It just took us some time because, you know, we're out here building cool things, we're we're the good guys. But then Facebook broke the world and we started thinking, huh? maybe we're not so good. it's been a transition of us moving into a reality. But tech has been a dream in this country for a long time. And it's just hard to wake up from it.

Gael Duez (41:13)

Yeah, that's a very good point because, know, the US situation is really specific with so much money poured by the oil industry to create this fog of misunderstanding around climate change, it's really a country that is specific regarding climate change. And from time to time, like, okay, you we've got environmental crises, not just an issue with climate change. So maybe in the US, you should refocus the fight, not against climate change. And my point is, if you focus on water consumption or air pollution, et cetera, eventually you will bump into the same issues, which is if we let main corporate actors treat common goods and the first one being nature and air and water as something that has no price, no value because it has no price and we cannot really seriously address these so-called negative externalities and reincorporate them in the system. I you can pick pretty much crisis that you want, you will get the same solution, like make them accountable, pay the right price for that we pay and that our children will pay, et cetera, et cetera. And this is why maybe in the US, the water, because water is something tangible, carbon is not. Plus carbon is related to energy and the US is a country cultivating the cult of energy. We are energetic people. We are people doing things, etc. mean, there is something almost philosophical here. Water is not. mean, water, you need water. Everyone can understand why you need water to grow a crop and to drink And the same with air pollution and noise pollution. So I was wondering if maybe you could get more traction with these fights that will actually help in any case, the fight against climate change and carbon emissions. seems regarding your answer that it really relates to the narrative framed by big tech companies. Like we are forced for good and we should get a free pass for everything. If I understood well

Masheika Allgood (43:26)

essentially articulated why I started Taps Run Dry. Because I had been having this existential crisis about why can't we move forward? And I realized you have to break it down to something tangible. No one's going to think, we're fine without water. Right? So like, that's why I chose water activism. That is specifically the thought process I went through what you just articulated, you have to make it something that is visceral to people. Carbon's not visceral. It just ask Al Gore, he's been beating this drum for like 30 years and it's not moving. It's just, it hasn't landed. So what will land? Water lands. And that's why I specifically sought out water as my area of activism. And then air pollution, same kind of thing. I live in a place where there is higher asthma on average than there is in other places. And we are a very financially viable area of the country to live in. Like people make money out here, but our air quality is horrible, right? that's something we should probably address. And it means something to all of the techies out here who are making massive amounts of money, but their children all have asthma. All right. That's a general statement, but a broad statement, that's why water and air quality have been where I'm hanging my hat.

Gael Duez (44:44)

Right. It's really an interesting approach. I've got the feeling that this is a very efficient way to get this little click in people's mind like, ⁓ actually, this is serious. I need to pay attention. And the moment they start to pay attention seriously about it, all the little pieces, they nicely add up. There is almost no U-turn. There is no turning back on the road to sustainability. Once you

you can pose because your mental health is at risk. Or, because your financial situation might force you to do so. But it's almost impossible to go back and say, no, actually, no climate change doesn't exist. And we don't have a water issue, as I thought 20 years ago. know.

Masheika Allgood (45:29)

It's going to be a battle, I think you're right. Once people get activated in this particular room, you can't let it go.

Gael Duez (45:37)

Well, Mashaika, thanks a lot for joining, for sharing this very specific and grounded fight regarding water consumption plus noise pollution and pollution as well. I've got the feeling that this is not the last time that we talk to each other.

Masheika Allgood (45:54)

I'd hope not.

Gael Duez (45:58)

Because there are these crazy stories you told me about NVIDIA and before, the ethics that was one of your fight before. I might also ask you to consider talking about this in very concrete ways with these concrete numbers that will be even more precise in a few months, because I think it will be a great talk at Green IO New York 2026 planets are aligned. I would love to see you on stage sharing your findings and your calculators and the way you actually hacked. love it. The Uptight Institute approach, like, hey, let's ask the engineers, but not like how much water do you consume? What are the safety requirements for you to get a good label? And then I'm going to be able to calculate things. Well, that's brilliant. I really love it. So anyway, thanks a lot for joining. I'm really happy that we kickstart the back to school season with this episode and hope to talk to you soon.

Masheika Allgood (46:55)

Yeah, thanks for having me. This was great. And yeah, I look forward to future conversation.

Gaël Duez (47:00)

Thank you for listening to this Green IO episode. Because accessible and transparent information is in the DNA of Green IO, all the references mentioned in this episode, as well as the full transcript are in the show notes. You can find these notes on your favorite podcast platform and of course on the website greenio.tech. If you enjoyed this interview, please take 30 seconds to give us five stars on Apple Podcast or Spotify. I know it's not easy to find a feature on these apps, but I trust you to succeed and give us some boost in their rankings. Sharing this episode on social media or directly with relatives working in the data center industry or simply being thirsty for information on the topic, I know I should stop doing these jokes. So sharing the episode is always a good idea to provide them with insights on this hot topic. You've got the point. Being an independent media, we rely mostly on you to get more responsible technologists on board. In our next episode, we will start exploring a bit the British digital sustainability landscape. After all, Green IO London is in five weeks. We will welcome back Anne Curie

who has investigated DeepSeek from a sustainability angle. Stay tuned. By the way, Green IO is a podcast and much more. So visit greenio.tech to subscribe to our free monthly newsletter, read the latest articles on our blog and check the conferences we organize across the globe. As I already mentioned, Green IO London is almost there. September 23rd and 24th are the dates. As a Green IO listener, you can get a free ticket to any Green IO conferences using the voucher GREENIOVIP. Just make sure to have one before the 30 free tickets are all gone.

I'm looking forward to meeting you there to help you, fellow responsible technologists, build a greener digital world.

❤️ Never miss an episode! Hit the subscribe button on the player above and follow us the way you like.

📧 Our Green IO monthly newsletter is also a good way to be notified, as well as getting carefully curated news on digital sustainability packed with exclusive Green IO contents.

#2191 RHONY Part One: Polly Want a Cackle?

Watch What Crappens

Justle's Glue and the Malfunctioning Elevator

The AI-powered Podcast Player