AI or Not

E037 - AI or Not - Masheika Allgood and Pamela Isom

• Pamela Isom • Season 2 • Episode 37

Share episode

Welcome to "AI or Not," the podcast where we explore the intersection of digital transformation and real-world wisdom, hosted by the accomplished Pamela Isom. With over 25 years of experience guiding leaders in corporate, public, and private sectors, Pamela, the CEO and Founder of IsAdvice & Consulting LLC, is a veteran in successfully navigating the complex realms of artificial intelligence, innovation, cyber issues, governance, data management, and ethical decision-making.

When you turn on your tap and nothing comes out, it's already too late. This eye-opening conversation with AI ethicist Masheika Allgood reveals the hidden environmental crisis lurking within our AI revolution, the massive consumption of drinking water by data centers.

Modern AI data centers are fundamentally different from those of the past. The specialized GPUs powering artificial intelligence generate extreme heat, around 190°F, requiring sophisticated cooling systems that consume millions of gallons of water annually. This isn't recycled or sea water, but drinkable water from the same aquifers that supply our homes and businesses. Most alarming? Many communities remain completely unaware, as tech companies frequently hide water usage behind non-disclosure agreements while openly discussing their electricity needs.

"Water is a finite resource in a way that nothing else on this planet is," explains Allgood, who developed the Data Center Water Consumption Calculator to empower communities. This free tool allows local officials and concerned citizens to estimate water impact before permits are granted. A typical 200-megawatt data center consumes approximately 2,000 acre-feet of water per year, potentially accounting for 10% of a local reservoir's capacity.

The crisis extends beyond traditionally water-stressed regions. Florida, despite its rainfall, faces water scarcity. Phoenix restricts housing development due to water limitations, yet continues approving data centers. Virginia struggles with similar issues. Since water governance happens at the local level, citizens have significant power to influence decisions through municipal meetings and advocacy with officials they know personally.

Have we considered whether technological advancement is worth sacrificing our most essential resource? Visit tapsrundry.com to access the Calculator and presentation materials you can share with local officials. Remember, when it comes to water scarcity, there's no backup plan. By the time you notice the problem, it may be too late to solve it.



[00:00] Pamela Isom: This podcast is for informational purposes only.

[00:27] Personal views and opinions expressed by our podcast guests are their own and not legal advice,

[00:36] neither health, tax, nor professional nor official statements by their organizations.

[00:43] Guest views may not be those of the host.

[00:48] Hello and welcome to AI or not, the Podcast, where business leaders from around the globe share wisdom and insights that are needed at this very moment to address issues and guide success in your artificial intelligence and digital transformation journeys.

[01:05] We have a unique and special guest with us today,

[01:09] Masheika Allgood.

[01:11] Masheika is founder of Ally Consulting and she's an AI ethicist.

[01:17] Masheika,

[01:18] thank you for joining me today.

[01:21] I certainly value you and your amazing credentials, which I'm going to ask you to talk about.

[01:28] But welcome to AI or Not.

[01:31] Masheika Allgood: Awesome. Thanks for having me. This was really cool. Ask. I haven't done a podcast in a long time and I tend to enjoy them, so I'm looking forward to it.

[01:39] Pamela Isom: How about you tell me more about yourself, your career journey.

[01:44] Tell me what's driving you to do what you do today because you do some, some very powerful things.

[01:50] And so I'd like to know more about that and what led you to where you are today.

[01:55] Masheika Allgood: Call it circuitous.

[01:57] I think that's the kindest way to put it.

[02:01] So I started out, I wanted to be a lawyer since the third grade. I saw the To Kill a Mockingbird and it just resonated. I understood the power dynamic between those with and those without at that age.

[02:15] And it was very powerful to me that you could use structure to make the powerless powerful,

[02:23] right? Like to shift the balance. And it just really resonated with me. And I've always kind of been motivated by that, the idea of the powerless just person,

[02:36] right, Being able to use the tools of the system to overcome the powerful unjust.

[02:42] So based on that, went to law school a couple times.

[02:47] So I have a Juris Doctorate. And then I went and got a Master's of law, which is a specialist degree in litigation and dispute resolution.

[02:55] Came out and wanted to do what I set out to do. But the business of law is very different than the theory of law,

[03:05] than the profession of law.

[03:07] And I landed in the business,

[03:11] was not none too pleased and so ended up going back to school.

[03:15] I had been teaching a course on business like White Collar Criminal Law. And I recognized that there were no lawyers or business people who had written the book. And I was like, how does that make sense?

[03:28] And I found that there's a,

[03:30] there was a space between law and business that wasn't well explored.

[03:34] So I went back and got a international business degree,

[03:37] focused really heavily on the tech side because I was always a techie, always liked it. It just wasn't a career. When I was a kid,

[03:45] not that I was aware of, finished school,

[03:47] I spent some time trying to get into tech in Florida, which at the time was not the thing.

[03:54] Came out to California,

[03:56] basically put all my eggs in this basket, all the way across the country. Didn't know anyone except for my sister.

[04:03] Came out here and basically worked my way up from the bottom. So started as a contractor,

[04:08] worked my way up to an FTE and ended up at Nvidia as an FTE doing something wholly non AR related.

[04:17] But when I get bored at a job, I learn new things.

[04:20] So I read this book,

[04:22] Mind at Play, about Claude Shannon and his work in information theory.

[04:28] Immediately started thinking that there are things that you can't transmit because there are feelings that you can't transmit to people unless they've had those feelings, right? And so I already understood intuitively that there would be some constraints.

[04:43] So then I started looking at how these theories were being used to impact legal processes.

[04:50] And I learned about the Compas algorithm,

[04:54] which is a sentencing algorithm that was being used basically to determine if people should stay in jail,

[05:01] if they should get longer sentences,

[05:03] things of that nature. And breaking down.

[05:06] I read a lot about how they broke down, the fairness when it came to the statistics.

[05:11] And that basically kind of put me on this AI ethics journey.

[05:15] So through that process, I founded the first AI ethics working group at Nvidia.

[05:21] And we spent a year trying to figure out how do we do AI ethics within the Nvidia context, within who we are or who we were in the world.

[05:33] Took that to the CEO.

[05:34] He wasn't interested.

[05:37] And that was pretty much my exit.

[05:40] And when I left, I knew that I had concerns around AI, but I knew that I couldn't really speak to the tech unless I understood it deeply. So I spent time really trying to understand how the tech worked.

[05:55] Took a whole lot of LinkedIn learning classes to actually understand, from a developer standpoint, how it worked.

[06:02] We really focused very heavily on natural language processing.

[06:05] And that was the foundation really of me getting deep into AI ethics. And I started Ally Consulting,

[06:13] basically to be able to share that knowledge to people who weren't techies. Like, I've been teaching in different capacities my whole life.

[06:21] So my skill set is to take the technical and highly specific and make it accessible to other folk. It's what I did as a product manager. I took business concepts and Made them make sense to engineers, engineering concepts made them make sense to business.

[06:36] So that just seems to be my wheelhouse. And ally consulting is basically my personal offshoot of that. When it comes to the AI ethics space.

[06:47] Pamela Isom: I like it and appreciate people that can take complex subject matter and really make it relatable so, so that we don't feel like things are over our heads or that is not applicable.

[07:00] I try to make,

[07:03] when I'm sharing examples or when I'm, when I'm talking to clients too,

[07:07] I try to make it hit home like I'm subject to talk about your family, your kids,

[07:16] anything, so that they realize that this is something that is, is relative to them.

[07:24] If it's a boardroom, for instance, then we're talking about the return on investment, we're talking about governance, different things, but it's important. And I appreciate people that can make technology and technology language, even legal language,

[07:36] make the jargon resonate.

[07:39] That's how you make impact.

[07:42] I appreciate that. Thank you for that.

[07:44] So now I want to know if we can talk about data centers.

[07:50] And I know we talked about this before.

[07:53] I want the listeners to hear what you have to say about data centers and resource impacts.

[08:02] And so let me say, I know your emphasis is on water scarcity, my emphasis is on energy consumption, but actually I care about both and I understand that this is due to complex workloads.

[08:16] Cloud computing is another contributor to this situation.

[08:20] But tell me more about data centers and water scarcity and your passion.

[08:25] Masheika Allgood: Yeah, I think so. I try to found these discussions by first saying that data centers for AI are not the data centers you grew up with.

[08:38] Masheika Allgood: Right.

[08:38] Masheika Allgood: They're not the same as a data center that you use for just processing data. Right. Like, so we grew up with data centers that were basically storage houses of, of data for different business purposes.

[08:51] Masheika Allgood: Right.

[08:52] Masheika Allgood: And that data needs to be stored, it moves through the system. But the when it comes to electricity and water isn't extremely high.

[09:03] Masheika Allgood: Right.

[09:04] Masheika Allgood: We've had those kind of data centers in existence for decades and they haven't really tipped the needle on anything.

[09:10] The reason data centers are part of the conversation now is in the last five or so years there's been this massive ramp up for data centers to support AI workloads.

[09:21] And the difference with an AI data center starts and almost ends with the chips.

[09:27] So the GPUs that you use for AI workloads process massive amounts of calculations in very short periods of time.

[09:37] And they do a lot of them.

[09:38] Masheika Allgood: Right.

[09:39] Masheika Allgood: Like the whole point of AI is it's just calculating way faster than any human, any other system we've ever created, right?

[09:48] So that's a lot of strain on the chip itself.

[09:52] These chips,

[09:53] because they process at such a level,

[09:56] run really, really hot.

[09:58] Like crazy, crazy hot. Like 190 degrees Fahrenheit hot, right? And that's each chip and these AI data centers have thousands of chips located in the same place. You have to cool all that.

[10:13] Like we all know if an engine runs too hot, it breaks down. The same thing happens with a gpu. If the chip runs too hot, it doesn't work as well and eventually it'll break down catastrophically, right?

[10:24] So you have to cool those chips. So the data centers we grew up with had air conditioning, right? You walk in, it was freezing,

[10:31] freezing.

[10:32] Pamela Isom: I'm talking freezing, right?

[10:34] Masheika Allgood: And they had like a cold aisle, but then there's a hot aisle where all the exhaust comes out. And that was crazy hot. And so basically it was, it was the temperature game, right?

[10:43] Push the heat here,

[10:44] pull the cold here,

[10:46] and that's how they kept it going. And that was sufficient for the kinds of chips and the kinds of workloads that we've been managing for, you know,

[10:53] the digital era.

[10:55] AI chips run so hot that that doesn't work.

[10:59] Like it's no longer sufficient.

[11:01] So you have to find different ways to manage that heat.

[11:04] And so now we're using liquid cool.

[11:07] And so you're running pipes through to pull that heat out through the liquid.

[11:13] And then you cool those liquids and then you run them back through. So it becomes this continuous loop of the pipe.

[11:21] Masheika Allgood: Right.

[11:22] Masheika Allgood: And they call these closed loop systems. So I wrote a substack article because the system isn't exactly closed.

[11:29] Like there is a system that goes through kind of like with coolant that, you know, cools the,

[11:34] pulls the heat, gets cooled, comes back around, pulls the heat. That system is closed,

[11:40] but you have to cool that system. And the second round of pipes is the open part, right? Because you basically take that heated up liquid and you pull it through a system that then pours water on it.

[11:55] And it's the water evaporation that actually cools down that liquid so it can go back in cold and come back in hot again, right? So the first loop is closed.

[12:04] Nothing's being lost from that system.

[12:06] But the heat exchange,

[12:08] the way that you cool that first loop is massively water intensive,

[12:13] right?

[12:14] So that's, that's where I, that's where the issue is when it comes to water is that second loop. But that second loop. So when you're cooling that water. It is.

[12:23] Through an evaporative process. Right. And so you're either pouring water on some sort of, like, sponge kind of material and that gets it cooled, or you're pouring it directly on, like, the pipe itself.

[12:35] But there's an evaporative process,

[12:38] and through that process, you're evaporating liquid.

[12:40] Masheika Allgood: Right.

[12:41] Masheika Allgood: Because the heat creates steam, evaporates liquid.

[12:44] There's also a fair amount that just kind of gets pushed out. Right. Because it's, you know, it's small droplets, lots of wind, and so you'll lose water that way. They call that drift.

[12:54] And then this system, as it goes through, it's just like an engine. The oil in your car, you put oil in your engine,

[13:02] it's really clear when it goes through,

[13:04] but after a certain amount of rotations, you pull it out and it's just like black and full of sediments.

[13:10] With these systems.

[13:11] Masheika Allgood: Right.

[13:12] Masheika Allgood: Same thing happens with these systems is at some point it gets full of gunk, like the. It's the.

[13:19] They have to treat the water to make sure it's not corrosive.

[13:22] Masheika Allgood: Right.

[13:22] Masheika Allgood: And at some point that starts to get, like, super concentrated. It ionizes. At some point that gets super concentrated. So then you have to get that water out of the system.

[13:32] And so that's.

[13:33] That's the second way that you lose water.

[13:35] Masheika Allgood: Right.

[13:36] Masheika Allgood: And so these processes basically take 1 to 2% of the water that's in that system is now completely lost.

[13:44] Masheika Allgood: Right.

[13:45] Masheika Allgood: Cannot use it again. You can try to treat some of it, but you're only going to get a percentage of it that's really water that's consumed.

[13:52] Masheika Allgood: Right.

[13:53] Masheika Allgood: And that's the concern.

[13:54] And when you think, oh, 1.5%, I mean, what is that?

[13:58] That's hardly anything.

[13:59] But if the system is running a million gallons of water a day and I'm losing 1.7%,

[14:06] then it's a problem.

[14:07] Masheika Allgood: Right.

[14:08] Masheika Allgood: So that's what we came up with, the data center water consumption calculator.

[14:14] And it essentially will tell you how much water needs to be in the system for it to run.

[14:20] And then that loss of water, that consumed amount, and then I can say this is how much is consumed in a day's operations, which lets me know this is how much will be consumed over the course of a year.

[14:32] Right.

[14:33] Now, mind you, that's looking at if it's running full tilt, and it's not always running full tilt, but a lot of times it is running full tilt. So the estimate is based off of that.

[14:42] And so when you were Talking about multi agents,

[14:46] those are things that bump up, like the amount of energy and electricity because you're running the GPU as hard as you can run it. Because there's so many systems that are interacting with each other.

[14:57] That kind of usage is going to cause you to run a bit hotter, right? You're going to use all of the electricity you've been allotted and then you're running full tilt.

[15:07] And that estimate that I'm putting out there is legit. Like that's your peak usage, but you're running peak usage more often than not, right?

[15:14] So with electricity, peak usage has two components. First, it's just the general electricity usage that you get through your, you know, public wiring.

[15:24] But second,

[15:26] you often have to use generators because at peak usage time for the computers is also peak usage time for humans. And so now you're running generators full tilt.

[15:36] And we know that generators,

[15:39] you know, they exude, you know, waste products, right? There's, you're running gas like it's an engine, an open engine. And so there are a lot of particles that get out.

[15:49] So you've got issues of air pollution with that.

[15:52] So I speak a lot about water, you speak a lot about electricity. But at the end of the day, our concern is community health.

[16:00] And without water,

[16:01] there is no community,

[16:03] right? And then if you're using,

[16:05] you know, generators,

[16:07] then you're negatively impacting community health even within a few short hours. But this is over months and years.

[16:15] It's going to have a square feet and bath.

[16:17] Pamela Isom: So the thing is, with electricity we have generators kind of as a fallback, but with water, what's our fallback? Like if we run out of water, what, what do we do?

[16:28] Masheika Allgood: Move.

[16:29] Like, that's, that's why I focus so heavily on water. There is no fallback for water. Like, you're right, with electricity, I can pull in a generator, I can use gas instead of coming through my wires.

[16:40] Like there are ways to make up for it. I can use solar battery packs. Like there are things that I can do. But when you run out of water, you run out of water.

[16:49] And the concern with water is this isn't reused water.

[16:54] This isn't salt water from the sea.

[16:57] This is potable water. This is water from your aquifer that is being used for us to drink.

[17:05] So it's competing with humans.

[17:07] And when that water is gone, it's gone for all of us, not just the machine.

[17:12] Pamela Isom: Okay?

[17:13] So I can't help but think when I listen to you how we are an innovative nation and how I Mean there ought to be some solutions, right?

[17:25] To address this problem. Like this seems like an old fashioned problem with I'm talking about keeping the data centers cool and keeping the chips cool.

[17:36] That seems like a problem that we often talk about how you want to think about your innovation and think about things to do to mitigate risk before those innovations become useless or before those innovations become ineffective towards society.

[17:54] And so what this conversation is bringing out is that the cost of that innovation is detrimental to society.

[18:04] And, and so we need to start looking at things from that angle,

[18:09] which is the ethical lens up front so that we think about this upfront and build this into the design of the chips or into the design of the solution centers earlier rather than later.

[18:25] And then when we do see this going on, do something about it.

[18:30] So I saw where I was reading where some of the companies like Microsoft and Google, different ones have set a goal for by 2030 they would be replenishing more water than they consume to help address this problem,

[18:47] particularly around water.

[18:49] And then I started looking at okay, well what are some things that we, they're looking at doing? And I saw where they were talking about locate data centers in cooler climates.

[19:00] Well, I don't know air cooling or liquid immersion cooling which you talked about,

[19:05] which kind of is,

[19:07] has a, has a cyclical effect. So it's not as effective as we would like air cooling could be if we, if we use air cooling properly, go back to the old way.

[19:17] But then we still have to figure out how to balance it with the types of chips that we're building. Right.

[19:22] And then use recycled or non potable water.

[19:26] And so I'm concerned with the quality of the water. I'm already concerned with the quality of the water that we drink.

[19:34] So because it's not where it should be. And so we've got to purify it and do all this stuff ourselves anyway.

[19:41] So I'm really concerned if the water is being recycled with and it starts to get the crud and stuff in it that you're talking about which makes it not even usable for cooling the data centers.

[19:55] So I'm concerned about that. So I want to hear from you. I said all that to let you know that I hear what you are saying and that I actually agree and started looking into what could be done to help mitigate the risk of this water stress.

[20:11] But I wanted you to talk more about your tool and you talked about it already. Some. But tell me more like why,

[20:20] who, who can use the tool, who should be using the tool and reiterate the point as to why we should be using the tool.

[20:28] Masheika Allgood: Okay, so I want to start first with some of your comments.

[20:30] Masheika Allgood: Right.

[20:31] Masheika Allgood: So the idea of water replenishment,

[20:35] the question that isn't being asked is, where are you putting that new water?

[20:40] Because it's not usually in the place where you took it from.

[20:44] Water is hyperlocal.

[20:46] So you can be globally water even, right? Like, oh, I replenished all the water I took out, but if you don't put it back in the community that you took it from,

[20:55] then you're still destroying that community.

[20:58] It's a big problem.

[20:59] Yes.

[21:01] You can't always replenish in the same way. Like, our technology around water replenishment requires certain things. Some places are harder to put the water back than others, depending on,

[21:11] like, how the aquifer is settled, how it's built. Like, it's. It's not just,

[21:17] I can take it out and put it back in. So when they talk about water replenishment, they're not giving you the details of I can't necessarily put it back in where I took it out.

[21:26] So that community that lost it may just be lost, but I gave it to somebody else.

[21:30] So overall, I'm equal. Like,

[21:34] that's the level of detail. Like any of these clients we're getting,

[21:37] when you get into the level of detail, it's not evening out the idea of placing all the data centers in cooler climates. Well,

[21:46] you know, the indigenous in Alaska, which is where they're trying to put these in Canada,

[21:51] they've all like, this is their land.

[21:53] So you're just gonna take their land. And again,

[21:58] to relocate data centers, like, so that's a. That's its own issue.

[22:03] Like, the colonial mindset of it. Like, oh, we're just going to take that land because they're not using it.

[22:08] Masheika Allgood: Right.

[22:09] Masheika Allgood: And then secondarily,

[22:11] as you know,

[22:12] these centers have to be located somewhere near where the data is going to be used.

[22:17] Otherwise there's other issues of trying to transmit that data across the country,

[22:22] across the world. So it's just not really technologically feasible if you're trying to have, like, very quick calculations for all your data centers to be located hundreds or thousands of miles away.

[22:34] So it just doesn't really even out.

[22:37] Masheika Allgood: Right.

[22:37] Masheika Allgood: Just briefly, just to. These are things that are being talked about. But how feasible are they?

[22:43] Masheika Allgood: Right.

[22:44] Masheika Allgood: So the reason why I created the data center calculator, consumption calculator, was specifically because water is a local concern.

[22:53] And I think for too long,

[22:56] this conversation has moved so fast that you can't have a real,

[23:03] like a municipality can't really have oversight or really a real stake in the negotiations if they don't know what's happening. So, like everything is about electricity. And so, you know, the electric company will come in and say, oh, we'll do this, that and the other.

[23:17] But I've yet to hear any conversation about people partnering with water,

[23:24] like with the county water district, the local water providers. You never hear that.

[23:28] Masheika Allgood: Right.

[23:29] Masheika Allgood: Electricity is very public, but you've got companies that are forcing local governments to sign NDAs, so they can't even speak about how much water they've been permitted to use.

[23:40] Right.

[23:40] So water is,

[23:43] it's a problem.

[23:44] Tech companies know that it's a problem and they've been hiding the information that's necessary for people to make real decisions.

[23:52] Masheika Allgood: Right.

[23:53] Masheika Allgood: And so that's basically why the calculator was created and those who should use it is anyone who is in a position to regulate or to permit or to zone a data center.

[24:05] Masheika Allgood: Right.

[24:05] Masheika Allgood: So before it's built,

[24:07] water needs to be at the table in the same way that electricity is. Because if you're building a nuclear power plant, it needs a lot of water. How are we not talking to the water folks before you get electricity to sign on to build a nuclear power plant?

[24:20] Like, there's a triumvirate that needs to be at this table that isn't right.

[24:25] Water is a finite resource in a way that nothing else on this planet is drinkable. Water is a finite resource. So you were talking about treatment and water treatments. Massively expensive.

[24:37] It's expensive to do, like to set up an entire system for it. It's not like tech companies couldn't do it,

[24:44] but it really changes their math on how much money they're going to get out of that data center or if it'll be profitable at all, given that AI workloads are not making the money that people were anticipating they were going to make.

[24:57] Like the major players are all losing money when it comes to these large language models. Like they're going off of funding, but like they're not revenue positive.

[25:07] So when you talk about now you need to add this huge infrastructure cost,

[25:12] they're going to balk at that. And so that has to be something that's required. They're not going to do it out of the just goodwill because it hurts the bottom line.

[25:20] But how can you require them to implement these systems if you can't even have an understanding of what they're actually going to cost you in terms of water?

[25:31] So I think you have to start with some level of data.

[25:35] And I was really frustrated for a long time. No one has any data. There is no data.

[25:40] And there was this whole conversation about, oh,

[25:43] you know, these, these things are so difficult to calculate. They have to be done at the data center level. And it's after it's built. Well, what's the point of calculating it after it's already been built?

[25:53] And they gave you an estimate and you just went with it like that's, that's not good business, that's not good governance. And so like our goal was to try to allow those in a position to be empowered to do better by their citizens.

[26:08] And so I think, you know, just having an initial estimate and then being able to have a real conversation about, well, what are. We have these assumptions in our tool.

[26:17] You know, how are our assumptions wrong? How is your estimate different based on your assumptions and really having a negotiation.

[26:25] Because at this time,

[26:27] basically companies come in and say, I need X amount of water. And then in two years they'll say, I need another 43% on top of that.

[26:34] And like, there's no structure to say no in a real way.

[26:39] So I think just bringing it to the forefront so people understand this is the same water that you need to drink. Like Phoenix is axing.

[26:50] Like, you can't build new housing in Phoenix because they don't have enough water for it. But Phoenix is a data center hotspot. It's a hub right now. People are like, there's a pipeline of people who want to build data centers in Phoenix,

[27:03] but they already know they don't have enough water to support future populations. So when do they get to say no?

[27:10] How do they get to signal like these kinds of conversations bringing this to the forefront so people understand,

[27:17] like they're not just anti AI or anti building good things for the community.

[27:23] Like, this is a real concern. We cannot replace our aquifers. When they're gone, they're gone and we have to protect them.

[27:31] I think bringing this into these kind of forums allows them not only the information they need through the calculator, but also just a little bit of public will to do the things that make the hard decisions that they haven't been enabled to do thus far.

[27:46] Pamela Isom: And better chip design, we could do better.

[27:50] Masheika Allgood: I mean,

[27:51] so I.

[27:53] My issue,

[27:54] the reason why I left Nvidia was it was very clear that there was only one person allowed to have ideas.

[28:03] And if it wasn't Jensen's idea, it just wasn't going to happen. But at Some point,

[28:08] you have to recognize that, you know, you are where you are because you got us there.

[28:15] But at some point,

[28:16] you have to be open to people who are seeing things in a different way.

[28:20] So I just. I cannot understand,

[28:24] given that Nvidia Engineering,

[28:26] it is the best engineering organization in the world. I mean, you can argue it,

[28:30] but it's an argument.

[28:33] They are undoubtedly in the top engineering organizations in the world.

[28:38] How have you not placed a challenge in front of that engineering organization to create the next generation of chips that runs cooler than the last generation of chips?

[28:49] It's a failure of leadership. I'm not going to put that on the engineering team. I know what they can do,

[28:54] but someone has to challenge them to do it. Someone has to give them that as a mandate. And for whatever reason,

[29:02] Jensen is unwilling or has been to this point, unwilling to focus on anything besides selling more chips.

[29:11] And that is not what he sold us on when we started with that company.

[29:15] There's a. There was an intro video,

[29:17] like, part of your orientation.

[29:19] And Jensen speaks in that video, and he says, we have a lot of competitors, but we have no peers.

[29:26] And that meant a lot to us at Nvidia. Like, we believe that we were doing something different, something better,

[29:32] that we had a focus that was more than just sell the next chip, that we were changing society in good ways.

[29:42] At some point around that 2020 mark,

[29:45] things changed and we became less of who we thought we were and more like everybody else.

[29:53] And it's to the detriment of the world, honestly,

[29:57] because that challenge would have changed everything. It still could,

[30:02] but that challenge would change everything. Because at the end of the day, it is the heat from the chips that is driving all of this.

[30:09] If those chips didn't run so hot, we wouldn't have this conversation because you wouldn't need all the extra power,

[30:17] all the extra water. You could cool them in a reasonable way. Like, at the end of the day, this is all driven by the heat from Nvidia's chips.

[30:26] So why is Nvidia doing anything about it?

[30:31] Pamela Isom: And then I also think about the data center designs.

[30:37] We could also.

[30:38] I'm not trying to solve it here, but there should be some more work done around the data center designs. Because I'm from the clean room area where we had the clean room where we walked around in these outfits so that particles could get here or there and just to give our hospitals and things.

[30:57] And so they're cold for a reason. But I understand that we've morphed away from that to some extent. But we probably need to think about that too.

[31:07] So I may do some research on that because I think it's both. It's data center design and it's the chip,

[31:14] of course, the root cause. I hear you saying what the root cause is. I want to go back to your water consumption calculator one more time.

[31:23] So who is, who can use it? Anybody. Can anybody get it?

[31:28] Masheika Allgood: It's, it's on my website. You just go in and you put how much electricity that the data center is going to use and it will immediately pop out all the numbers you need to know as a general estimate of how much water we'll need to run the system and then how much will be consumed at a daily,

[31:45] monthly, yearly output. Right now it's in gallons and liters. I'm going to add acre feet because acre feet is what translates directly to your reservoir.

[31:56] Discuss reservoirs in terms of acre feet of storage, which is something like 35, a little over 35,000 gallons of water.

[32:03] So just doing that last conversion so that you can understand if you've got a 20,000 acre feet reservoir and you're putting in a 200 megawatt data center, that 200 megawatt data center is going to take about 2,000 acre feet of water per year.

[32:21] It's going to consume about 2,000 acre feet of water per year. You might want to rethink that,

[32:27] given the size of your reservoir, because you're not necessarily going to be able to replenish that much water reliably every, every year through rain and whatnot. And that's, that's 10% of your reservoir just going to data center operations.

[32:42] You've got to support people,

[32:44] businesses,

[32:46] possibly agriculture and other purposes. So just having that as a starting point for your understanding, I think really allows you to better plan your city operations going forward.

[32:59] Pamela Isom: Okay. So if you have mentioned to me about local action and what there needs to be some local action to support what's really needed,

[33:10] I think that I can see and understand what you're saying, but I'd like you to elaborate a little bit more. So I know there's the citizens that can take local action,

[33:21] then there's our political officials and our city council, the municipalities, et cetera, et cetera. Can you elaborate what you mean by that? And what, what can we do?

[33:31] Masheika Allgood: Okay, so essentially a data center doesn't just appear in a city.

[33:37] Masheika Allgood: Right.

[33:38] Masheika Allgood: It goes through a pretty extensive process when it comes to zoning and permitting.

[33:42] Masheika Allgood: Right.

[33:43] Masheika Allgood: So they have to go and get a permit for water. Like they have to Ask the city or the county or whoever manages the water in that region,

[33:51] can I withdraw X amount of water per unit?

[33:55] Masheika Allgood: Right.

[33:56] Masheika Allgood: That's a process.

[33:57] They also have to go through and get usually some sort of special zoning variants to actually build the data center where they want to build it because it's kind of quasi industrial,

[34:07] a lot of times they're trying to build it kind of close to housing,

[34:10] which is not a good look because noise pollution,

[34:13] air pollution, light pollution.

[34:15] So there's a lot of issues that go on with that. But you know, data center is not like an easy zone.

[34:21] Masheika Allgood: Right.

[34:22] Masheika Allgood: Because the data centers we used to build were of kind of warehouse size. The data centers they're building for AI are campuses.

[34:30] They're massive. And so they all need some sort of special zoning permit.

[34:34] Masheika Allgood: Right.

[34:35] Masheika Allgood: In that process there's citizen action that can happen.

[34:39] Masheika Allgood: Right.

[34:39] Masheika Allgood: Like I as a citizen can go to these meetings where, you know, these questions are coming up and my council is going to vote and I can speak on it and say I am concerned for these reasons.

[34:49] Have we considered the water?

[34:51] I don't see anything.

[34:52] No one's talking about water. Is this going to impact my water? Particularly if you're in some sort of agricultural district or you're in a district where water is important either for recreational purposes or something else.

[35:05] Masheika Allgood: Right.

[35:06] Masheika Allgood: You know, you can go and have a conversation with their local leaders or even at the meetings themselves. I went and spoke at my local water district meeting. They gave me a minute and a half,

[35:15] Right.

[35:16] I spoke for my minute and a half and I submitted the Taps run Dry presentation which is also on my website as part of the meeting minutes.

[35:25] Masheika Allgood: Right.

[35:25] Masheika Allgood: You can do that as well. But it just, it's something that your local leaders will now read because it's part of the minutes.

[35:31] Masheika Allgood: Right.

[35:32] Masheika Allgood: So that's your, your first kind of citizen action as far as like local and government leaders. Once again, if you're sitting on those boards,

[35:40] this is your action is to ask those difficult questions like you're coming to me asking me for X amount of water,

[35:47] what is your estimate versus my estimate?

[35:50] Like you know, how you know. This is the calculator I used.

[35:54] These are the assumptions in that calculator. Are you saying you're not going to run at peak usage365? Well, how often do you run at peak usage based on your other data centers?

[36:03] Masheika Allgood: Right.

[36:03] Masheika Allgood: Have an informed conversation.

[36:05] I think there's also a real need for us to rethink our water governance policies.

[36:13] Like right now, if you over withdraw your allotted amount There aren't really any like mechanisms to punish you for.

[36:23] Like if you go through and look, maybe your state is different. I live in California and I've done research in other states and I'm not seeing that there's mechanisms to punish people who over withdraw.

[36:33] There's no accountability.

[36:35] Yeah, there's no accountability.

[36:36] Also if like the way that permits are given is if an aquifer is running low,

[36:43] they won't give new permits.

[36:45] But the old permits don't get touched.

[36:48] So I can, if I have a permit to withdraw water, I can withdraw it to the ground.

[36:53] Like I can withdraw it until it's done if I already have a permit. There's no way to modify permits of people who have been given permits.

[37:01] You just can stop new people from getting permits.

[37:04] So that's pretty terrifying, right? Like once you're in, you're in. But this is the system. So I think there is space for us to have real conversations.

[37:13] And that's at a local level because water is a local concern. The state doesn't really decide in most states, it doesn't really decide how water is permitted or governed within the localities.

[37:26] It's a,

[37:27] it's a thing within the US based on agriculture and ranch in our history that it's hyperlocal when it comes to water. And people have fought hard for generations to keep it that way.

[37:39] And I don't see that changing anytime soon. Which means as a citizen you have real power to really impact how these situations work because they're at your level.

[37:50] This person who you see at the grocery store every day is making these decisions. And there's no big, big nameless, faceless mass you know, that can come in and overruled them.

[38:00] Like this is truly local power when it comes to women.

[38:05] Pamela Isom: So is it in our best interest to get to know the city officials,

[38:13] the county officials?

[38:15] It's at that level. Right? City, county, but not at state level.

[38:20] And then in the.

[38:23] Masheika Allgood: Yeah, city, county. And then there's water, special water management districts. So just special groups that, that get to manage water. Like that's when it comes to water, it's all at that very hyper local level.

[38:34] So there are at some, like there are things you can try to get passed legislation at the state level. But if you want to have immediate impact,

[38:43] because you can.

[38:45] It's all at your city counter, county,

[38:48] local water management district.

[38:50] Pamela Isom: Okay,

[38:51] okay,

[38:52] that's good. Okay. So now we know,

[38:55] now we've gotten educated.

[38:57] Now we know what we can do. And then I also got enlightened on Seriousness about the implications because the water will run dry. And what's the name of the talk?

[39:11] Taps Run Dry.

[39:12] Masheika Allgood: Is that what it's called, Taps Run Dry? I created a platform, so there's a website, there's a foundational presentation that I created that you can give to your local leaders.

[39:23] And we're doing a lot of work within that space.

[39:26] But it's specifically for municipal level action to educate our leaders, for us to take our own, like to get involved.

[39:34] But that's when the taps run dry. They run dry.

[39:38] Pamela Isom: Are we nearing that and are certain locations nearing that more than others?

[39:44] Masheika Allgood: Yes,

[39:45] but it's very difficult to give like at a more like national level. It's difficult because border is hyper local.

[39:55] The data sharing isn't quite the same from state to state, city to city.

[40:00] So there's no really cool fed sites you can look at.

[40:03] Some states provide more information than others,

[40:06] but Phoenix is in a bad way.

[40:10] You know, border scarcity is an issue all over the country. Like people tend to think, oh, it's only a problem on the west coast,

[40:18] but Virginia is having problems.

[40:20] Florida is doing wastewater treatment for drinking water. Like I'm from Florida and we get rained on all the time. There's water scarcity in Florida.

[40:31] Like the entire US has issues of water scarcity in literally every state.

[40:37] So it's not something that's just a concern.

[40:40] If you live in a dry state, you can live in a pretty wet state.

[40:44] And water is still an issue right now. Like the world is changing.

[40:49] You're not getting rain where you used to get it, you're not getting it to the intensity that you used to get it. So all of that impacts and we're drawing way more water.

[40:58] We've oversubscribed our aquifers all around the country,

[41:03] so we're withdrawing way more than the. Thank you. The earth can replenish.

[41:08] It takes a long time for that water to get down in a lot of these aquifers. Like if you have a reservoir that works better. Like the water gets there quicker, but it also evaporates quicker.

[41:19] And given that we're having much hotter periods, much hotter intensity to those heat periods, evaporation is a real concern.

[41:27] So water is being attacked just by the nature of the climate that we live in.

[41:33] And we're over subscribing it. That math doesn't matter.

[41:37] Pamela Isom: Well,

[41:38] that was very informative.

[41:40] First of all, is there anything else you want to share with me before you share either words of wisdom or a call to action or something that you want us to take, to walk away knowing and understanding in addition to what you've already talked about.

[41:55] But is there anything else before that?

[41:59] Masheika Allgood: So I think,

[42:00] I think what I want people to walk away with is the knowledge that water impacts all of us.

[42:05] It is not something that we can expect somebody else to deal with. It is going to impact all of our lives.

[42:13] And there is no one coming to say this.

[42:16] So we often in the US wait for someone else to handle a problem,

[42:21] but we don't have the luxury when it comes to this.

[42:25] And I also want people to understand you don't have to become a water warrior.

[42:29] You don't have to,

[42:30] you know, take off of work. And I mean, you can call your congressman during your lunch break. Not congressman, your city councilman during your lunch break.

[42:39] Masheika Allgood: Right?

[42:40] Masheika Allgood: You can share this presentation in an email like while you're home at night. Like there are ways to be active without having to go full activist.

[42:52] But I just want people to understand that this affects them in a real way.

[42:58] And you're not going to feel it until it's too late.

[43:02] Like when you get to the point where you turn on your tap and it doesn't work. And that is just terrifying. We've not done the math in this country,

[43:11] really in any locality to understand how much water we have left in our individual aquifers and what that means in terms of the amount of time we can support the people that are in our cities.

[43:24] Like we just haven't done that math. And there's no one putting pressure on us doing that math.

[43:29] So don't assume because you haven't heard anything, everything is fine.

[43:34] You need to get ahead of this because if you don't, by the time you get into it, it's just going to be too late.

[43:41] Pamela Isom: Okay, well, I think that those were really good words of wisdom. And also I hear a call to action in there.

[43:47] So I will allow the listeners to get back to me and let me know kind of what they're thinking about. Everything that we've said here today,

[43:58] what you've said today, I'll allow the listeners, they sometimes get back to me and let me know how this impacted them. So I'm going to be listening for that and as I hear things, I will let you know.

[44:07] But our call to action is let's get the tool,

[44:11] get our hands on the tools to better understand how to use it and how we can share it.

[44:18] If our local community leaders aren't using it,

[44:21] help them understand how they can use it to better understand how water is being projected to be utilized in the environment.

[44:31] So I think that that's really helpful.

[44:34] And then I would encourage you to keep going.

[44:41] Good job.

[44:42] Masheika Allgood: Thank you. I appreciate it.

[44:43] Pamela Isom: Yeah.

[44:44] Masheika Allgood: Thank you for this conversation. I enjoyed it.