WEBVTT - Ep. 902: The Pros, Cons, and Complications of Data Centers 0:00:08.760 --> 0:00:13.320 If this is the me Eater podcast coming at you shirtless, severely, 0:00:13.480 --> 0:00:18.560 bug bitten, and in my case, underwear listeningcast, you can't 0:00:18.600 --> 0:00:22.920 predict anything brought to you by first Light. When I'm hunting, 0:00:23.079 --> 0:00:27.640 I need gear that won't quit. First light builds, no compromise, 0:00:27.800 --> 0:00:31.480 gear that keeps me in the field longer, no shortcuts, 0:00:31.720 --> 0:00:34.839 just gear that works. Check it out at first light 0:00:35.000 --> 0:00:38.640 dot com. That's f I R S T L I 0:00:38.760 --> 0:00:45.000 T E dot com. Okay, folks, here, by popular demand 0:00:45.320 --> 0:00:48.720 is our show where we dive into data centers from 0:00:48.720 --> 0:00:52.240 a natural resources perspective. People are always writing in asking 0:00:52.240 --> 0:00:53.479 about why don't we cover this more? 0:00:53.560 --> 0:00:55.120 We should do it. Part of the. 0:00:55.080 --> 0:00:57.520 Problems I was trying to find the perfect person to 0:00:57.600 --> 0:01:00.000 do it, because I just want someone that can look 0:01:00.160 --> 0:01:05.000 at this at the data center debate from an informed perspective, 0:01:05.160 --> 0:01:07.960 from an even perspective, someone that's not trying to drive 0:01:08.040 --> 0:01:10.520 a particular narrative or that's in the pocket of a 0:01:10.520 --> 0:01:13.120 certain set of interests, like someone to take a journalistic 0:01:13.200 --> 0:01:15.480 look at what is going on with data centers. And 0:01:15.800 --> 0:01:19.960 she's here and Davis Vaughan and was a longtime Wall 0:01:19.959 --> 0:01:24.000 Street journal Energy and Finance reporter. Then she became an 0:01:24.000 --> 0:01:27.760 investment analyst, which is basically where you like do investigative 0:01:27.800 --> 0:01:29.679 reporting on companies. 0:01:30.000 --> 0:01:33.320 That's right. I was like an investigative reporter for portfolio managers. 0:01:33.480 --> 0:01:35.680 Got it so if they wanted to make an investment, 0:01:35.760 --> 0:01:37.800 you go dig up the dirt or they. 0:01:37.880 --> 0:01:40.080 Get the best version of the truth you can find. 0:01:39.920 --> 0:01:41.240 Best version of the truth you can find. 0:01:41.880 --> 0:01:46.120 And now Anne has re entered journalism and is writing 0:01:46.120 --> 0:01:52.040 a book which likely likely will be called Gigawatt, and 0:01:52.120 --> 0:01:55.760 it is about the AI industrial boom, how it is 0:01:55.800 --> 0:02:02.559 reshaping the energy industry, the industrial landscape, and the American heartland. 0:02:03.240 --> 0:02:05.800 She currently writes this is this is where I found Anne. 0:02:07.280 --> 0:02:10.960 She currently writes a column for the Information So it's 0:02:11.600 --> 0:02:15.680 information dot Com, the Information dot Com, the Information friends 0:02:15.680 --> 0:02:18.400 send me, friends would send me your columns. And she 0:02:18.480 --> 0:02:23.120 writes a column on AI infrastructure for the Information dot Com. 0:02:23.280 --> 0:02:26.000 That's a publication based out of Silicon Valley. And what 0:02:26.080 --> 0:02:28.760 I liked about it about your writing when I would 0:02:28.760 --> 0:02:32.359 read it is it was like it was very refreshing 0:02:32.480 --> 0:02:36.320 to hear someone looking at the issue that I felt 0:02:36.800 --> 0:02:40.040 hadn't already arrived at all. The answers I was just 0:02:40.240 --> 0:02:43.760 asking questions and also taking, as we're going to get into, 0:02:44.360 --> 0:02:49.840 taking a lot of measurements, measurements of water, measurements of energy, 0:02:49.880 --> 0:02:53.720 measurements of footprint, and putting those measurements sort of into 0:02:53.760 --> 0:02:57.640 context and comparing them to things that people might understand. 0:02:58.720 --> 0:03:03.440 So rather than hearing about some kind of wattage, comparing 0:03:03.760 --> 0:03:06.960 that to other industrial applications, other things that we're all 0:03:07.000 --> 0:03:10.960 familiar with, so you start to form a relative picture. 0:03:12.000 --> 0:03:13.440 So thank you for coming on, and thank you for 0:03:13.480 --> 0:03:15.600 sharing a bunch of your reporting, especially because you're doing 0:03:15.639 --> 0:03:18.000 this before your book comes out, which is which is 0:03:18.000 --> 0:03:18.400 good of you. 0:03:18.600 --> 0:03:21.080 Yeah, well, thank you. I'm excited to be here. I'm 0:03:22.040 --> 0:03:26.919 excited about your podcast and having come from the Blue 0:03:27.000 --> 0:03:30.680 Ridge Mountains and love the outdoors. It's a big part 0:03:30.720 --> 0:03:32.840 of sorting through what's happening today. 0:03:32.960 --> 0:03:33.600 Got it right? 0:03:33.840 --> 0:03:37.200 Can you start out by this kind of a dumb 0:03:37.280 --> 0:03:38.880 question about this I want, I want to do it. 0:03:39.520 --> 0:03:44.440 Can you start out by like telling people in twenty 0:03:44.600 --> 0:03:48.440 twenty six, when they see data center in a headline 0:03:48.480 --> 0:03:53.120 in the newspaper, Okay, what exactly. 0:03:52.720 --> 0:03:54.080 Are they talking about? 0:03:54.520 --> 0:03:57.200 It's a dumb question of just seting what is a 0:03:57.400 --> 0:03:58.800 data center today. 0:03:58.880 --> 0:04:00.920 Well, I'll tell you what it is not. It is 0:04:01.000 --> 0:04:05.080 not clouds in the sky. The data center is referred 0:04:05.120 --> 0:04:07.480 to as the cloud. It's like the I think it's 0:04:07.520 --> 0:04:11.680 one of the best marketing decisions tech has ever made 0:04:11.720 --> 0:04:15.000 for itself is to call data centers a cloud. These 0:04:15.200 --> 0:04:19.000 are in the cloud. It's in the cloud, and this 0:04:19.920 --> 0:04:26.000 cloud is really an industrial complex and so it's an 0:04:26.000 --> 0:04:31.520 industrial scale facility back kind of pre AI when we 0:04:31.680 --> 0:04:34.720 just thought of it as the cloud that holds our 0:04:34.920 --> 0:04:39.480 entire digital lives, and that means everything that you know, 0:04:39.880 --> 0:04:42.320 we waste our time on, and everything that is vital 0:04:42.360 --> 0:04:46.279 to our society. So you know, how you save your pictures, 0:04:46.279 --> 0:04:48.839 how you strain your movies, how you listen to your podcasts, 0:04:49.320 --> 0:04:52.080 how you get money out of the ATM, you know 0:04:52.120 --> 0:04:56.000 how your payroll gets paid, how you know life saving 0:04:56.040 --> 0:04:59.880 machines stay on and scan your body and track data. 0:05:00.560 --> 0:05:08.239 It's all that and the you know, cloud. Data centers 0:05:08.279 --> 0:05:12.960 have been a pretty big part of the digital landscape 0:05:13.200 --> 0:05:16.400 since you know, kind of the mid two thousands. A 0:05:16.440 --> 0:05:20.680 lot of companies used to have basically their own data rooms, 0:05:20.680 --> 0:05:25.240 computer rooms, and they'd have servers that were in a closet, 0:05:25.600 --> 0:05:29.000 and eventually some of those got bigger. And of course, 0:05:29.360 --> 0:05:33.320 since you know, for decades we've had government supercomputers that 0:05:33.360 --> 0:05:36.240 were far bigger than that. But this idea of a 0:05:36.440 --> 0:05:41.560 data center that started to grow commercially came with the 0:05:41.640 --> 0:05:46.599 rise of the Internet. These these buildings, you know, would 0:05:47.360 --> 0:05:51.960 uh serve a variety of different customers. And the notion 0:05:52.200 --> 0:05:56.799 of these big tech software companies becoming cloud giants really 0:05:56.839 --> 0:06:04.280 meant that Google, Microsoft, Amazon began to build buildings with rows, 0:06:04.600 --> 0:06:09.760 hallways of chips and servers and all the equipment that 0:06:09.880 --> 0:06:13.880 keeps them at this goldilocks you know, environment and temperature 0:06:15.800 --> 0:06:18.919 and fiber would go in and out and and data 0:06:19.000 --> 0:06:22.640 would you know, Uh, it was very important to have 0:06:22.720 --> 0:06:25.600 data have what is called low latency, so they tended 0:06:25.600 --> 0:06:31.760 to be in urban centers and the low latency low 0:06:31.800 --> 0:06:34.839 latency means quick like you know, you know when you 0:06:34.880 --> 0:06:37.520 when you remember the days when you would try to 0:06:37.600 --> 0:06:40.400 watch a video and it would say buffering, buffering, Yeah, 0:06:40.400 --> 0:06:45.920 that's low latency. And so you need instantaneous data for 0:06:46.000 --> 0:06:49.480 something life saving, like you know, an emergency response system. 0:06:50.080 --> 0:06:54.359 Traders want instantaneous data. We get upset if we can't 0:06:54.480 --> 0:06:56.760 you know, if a text doesn't come through right away. 0:06:58.680 --> 0:07:01.200 There are certain businesses that really are mission critical and 0:07:01.240 --> 0:07:04.719 that need to be near a city. So a lot 0:07:04.760 --> 0:07:09.680 of data centers got built in smaller fashion before the 0:07:09.720 --> 0:07:14.880 AI age, in especially northern Virginia near Washington, d C. 0:07:15.600 --> 0:07:20.800 And then some big cities like Dallas, Chicago, Phoenix out 0:07:20.840 --> 0:07:24.080 on you know, the West coast and both the Silicon 0:07:24.160 --> 0:07:28.120 Valley area and then up in the Pacific Northwest near Seattle. 0:07:28.760 --> 0:07:33.200 Was it was it really that proximity in those days, 0:07:33.200 --> 0:07:37.560 which is just like yesterday. Yeah, the proximity to you 0:07:37.960 --> 0:07:42.320 to user groups was important, like physical proximity to big 0:07:42.440 --> 0:07:43.600 populations of people. 0:07:44.560 --> 0:07:48.840 Yes, because the there's a limit to how quickly data, 0:07:49.480 --> 0:07:54.720 you know, will travel on fiber. And we're talking, you know, 0:07:55.160 --> 0:07:59.440 just nanoseconds. We're not talking, you know, we're we're really 0:07:59.520 --> 0:08:03.160 unwilling to wait for a few seconds for something to 0:08:03.200 --> 0:08:06.080 come through. So they were built under that, you know 0:08:06.960 --> 0:08:11.400 premise that giant amounts of fiber got built in the 0:08:11.440 --> 0:08:14.760 dot com era, there was a whole bunch of it left, 0:08:15.520 --> 0:08:17.720 you know, didn't get used, and then we kind of 0:08:18.080 --> 0:08:21.680 built into that in the two thousands and in the 0:08:21.720 --> 0:08:27.080 twenty tents, and when COVID happened, people started using online 0:08:27.320 --> 0:08:31.920 digital resources like crazy, and even more continued to be built. 0:08:32.440 --> 0:08:37.200 And so then when chat, GPT and AI as a 0:08:37.200 --> 0:08:40.320 new paradigm kind of really came on the scene around 0:08:40.320 --> 0:08:46.040 twenty twenty two, there began to be the conception of 0:08:46.080 --> 0:08:49.880 a whole nother kind of scale of data center. So 0:08:50.320 --> 0:08:56.000 AI data centers are bigger, more power hungry, more resource hungry, 0:08:56.040 --> 0:09:03.880 really because AI models operate on a scaling phenomenon, and 0:09:03.960 --> 0:09:07.360 so the idea is that it's not just a building 0:09:07.400 --> 0:09:12.160 with a whole bunch of different companies siloed computing taking place, 0:09:12.160 --> 0:09:16.240 but instead it is one big campus where all the 0:09:16.360 --> 0:09:19.800 chips are connected to all the other chips and they're 0:09:19.800 --> 0:09:22.480 all talking to each other, and it's like one big brain. 0:09:23.080 --> 0:09:28.320 And the more powerful models that have been trained have 0:09:28.440 --> 0:09:32.200 operated on this concept, and so we're building even bigger 0:09:32.200 --> 0:09:37.560 ones now and that they operate all together big that 0:09:37.720 --> 0:09:42.240 means bigger footprint for these data centers, more industrial equipment 0:09:43.240 --> 0:09:46.120 to bring them the electricity and to cool them. And 0:09:46.160 --> 0:09:48.160 that's why they are basically factories. 0:09:48.720 --> 0:09:52.319 And all the time that we've been like, well, let 0:09:52.320 --> 0:09:55.200 me reapproach that maybe you can correct me on the timeline, 0:09:55.240 --> 0:09:58.000 but at what point was it a year ago? 0:09:58.679 --> 0:09:59.680 Was it eighteen months ago? 0:09:59.760 --> 0:10:01.640 At what point was it that all of a sudden 0:10:01.679 --> 0:10:05.959 this debate about data centers and the environmental impact of 0:10:06.040 --> 0:10:10.560 data centers exploded. You probably have a different view of 0:10:10.559 --> 0:10:12.440 it because you follow this and report on it, But 0:10:12.480 --> 0:10:15.520 try to, like in your mind, think of think of 0:10:15.559 --> 0:10:19.760 just a normal, concerned citizen who reads a few newspapers 0:10:19.800 --> 0:10:23.240 in the morning. Like, help me, remember when was it 0:10:23.240 --> 0:10:25.199 at all sudden? It was just like the issue. 0:10:25.400 --> 0:10:28.600 It was that issue around the end of twenty twenty 0:10:28.640 --> 0:10:30.320 five and early twenty twenty six. 0:10:30.360 --> 0:10:33.960 Okay, so it was recent, yeah, Because what struck me 0:10:34.080 --> 0:10:38.080 is where was the argument about data centers prior to that? 0:10:40.640 --> 0:10:41.120 Was there one? 0:10:42.520 --> 0:10:45.360 Yes, there was, And if you go back you will 0:10:45.400 --> 0:10:50.360 find some pushback from a few years ago. I think 0:10:50.440 --> 0:10:55.720 that the scale just had not hit people. And so 0:10:56.040 --> 0:11:01.880 this AI boom has gone through several stages, and you know, 0:11:01.920 --> 0:11:04.959 at first it was just like, what is this AI thing? 0:11:05.120 --> 0:11:08.360 What is chat GPT? It's going to change everything? And 0:11:08.400 --> 0:11:11.000 then there was this stage when I was actually still 0:11:11.000 --> 0:11:14.160 working as an investment analyst where everyone was saying how 0:11:14.240 --> 0:11:16.719 much power are we going to need and where is 0:11:16.760 --> 0:11:20.480 it going to go? And you know, the financial and 0:11:21.040 --> 0:11:25.920 energy and tech types were mobilizing and everyone was saying, 0:11:26.480 --> 0:11:29.600 can we really use this much power? Then there was, 0:11:29.640 --> 0:11:33.240 like maybe in the fall of last year, an absolute 0:11:33.320 --> 0:11:37.040 panic over whether AI was a bubble and whether we're 0:11:37.040 --> 0:11:39.960 spending too much because at that stage everyone was really 0:11:40.000 --> 0:11:42.800 focused on how much the tech guys were going to 0:11:42.840 --> 0:11:46.079 be spending and you know, every quarter and how they 0:11:46.080 --> 0:11:48.440 had gone from just being so cash rich to spending 0:11:49.000 --> 0:11:51.120 you know, a great deal of their cash flow. And 0:11:51.200 --> 0:11:57.680 so all during that time, the the the you know, 0:11:57.840 --> 0:12:01.240 wildcatters that were trying to put AI together and find 0:12:01.320 --> 0:12:05.720 places to put it were going to communities and starting 0:12:05.760 --> 0:12:12.600 to make plans for data centers and individual communities were 0:12:12.640 --> 0:12:16.600 having meetings and saying, what do we think about this? 0:12:18.280 --> 0:12:22.760 But the populace in general is not super engaged on 0:12:23.120 --> 0:12:27.480 you know, their local community level. There's we've kind of 0:12:27.520 --> 0:12:31.240 turned polarized and less local news to consume anyway, and 0:12:31.360 --> 0:12:35.400 the digital age sadly, and and you know, people have, 0:12:35.920 --> 0:12:38.480 you know, whatever they're concerned about on the national level, 0:12:38.520 --> 0:12:44.200 and all of a sudden, these campuses started to get 0:12:44.200 --> 0:12:48.680 constructed and people started to say, wait a minute, like, 0:12:48.840 --> 0:12:53.160 I didn't know this was happening, and well, you know, 0:12:53.240 --> 0:12:55.800 some of some of this may have been done just 0:12:55.880 --> 0:12:58.719 like any kind of industrial real estate project might have 0:12:58.760 --> 0:13:00.679 been in the past, where people go to the local 0:13:00.800 --> 0:13:04.640 zoning authorities they say, Hey, we're with this company, we 0:13:04.679 --> 0:13:07.440 want to buy this land. Can we put this application in? 0:13:08.880 --> 0:13:11.720 You know, normally the public is not watching every move 0:13:12.120 --> 0:13:15.800 on that front, but all of a sudden, there was 0:13:15.840 --> 0:13:20.680 this collective realization that there have been a kind of 0:13:20.840 --> 0:13:25.640 rush out into the middle of the country because these 0:13:25.720 --> 0:13:28.640 data centers were on a totally different scale than the 0:13:28.640 --> 0:13:31.120 cloud data centers, and we're going to need to go 0:13:31.640 --> 0:13:34.439 places where there was enough land or where there was 0:13:34.600 --> 0:13:38.560 enough power, where there was some headroom in you know, 0:13:39.520 --> 0:13:42.840 whether it was water, the grid, and. 0:13:44.400 --> 0:13:44.800 That is. 0:13:45.720 --> 0:13:47.559 It kind of started to reach a fever pitch. I 0:13:47.559 --> 0:13:50.319 would say, I noticed it on a trip to Wisconsin 0:13:50.880 --> 0:13:59.200 in early February, where the backlash had grown just exponentially. 0:13:59.360 --> 0:14:02.080 Yeah, I want to do a thing where you give 0:14:02.120 --> 0:14:04.599 me you can pick, or you can do neither, or 0:14:04.640 --> 0:14:06.960 we can play a game to figure out who does it. Like, 0:14:07.000 --> 0:14:08.920 I want to do the positives. Just before we get 0:14:08.920 --> 0:14:10.400 into some of the details. I want to do the 0:14:10.440 --> 0:14:12.800 positives and negatives. You can take both of them. I 0:14:12.840 --> 0:14:16.440 can take one, you can do both. You can sell 0:14:16.480 --> 0:14:19.440 me on the hysterical view. I don't want to load 0:14:19.440 --> 0:14:22.320 it too much. You can sell me on like why 0:14:22.360 --> 0:14:25.360 this is upsetting the people, and you can sell me 0:14:25.440 --> 0:14:28.840 on why it's exciting to some right, or we can. 0:14:28.760 --> 0:14:30.840 Split the job up to you. 0:14:31.280 --> 0:14:33.880 Well, okay, So the worries are that data centers are 0:14:33.920 --> 0:14:37.360 going to destroy our environment, They're going to you know, 0:14:37.520 --> 0:14:42.600 destroy our health, That they'll drain Lake Michigan, they'll you know, 0:14:43.000 --> 0:14:47.240 raise the temperature above the buildings and into our neighborhoods 0:14:47.280 --> 0:14:51.320 so poison the water supply. They'll spike our power bills, 0:14:53.200 --> 0:14:55.920 they'll only take subsidies and not give anything back to 0:14:56.480 --> 0:15:01.040 you know, local communities, and and that it's all happening 0:15:01.040 --> 0:15:04.720 in secret. So I think that's part of the that's the. 0:15:05.200 --> 0:15:07.160 That's the vibe I get. That's the vibe I get 0:15:07.160 --> 0:15:07.680 of the panic. 0:15:07.760 --> 0:15:10.600 Yeah, yeah, yeah. And and not to even mention that 0:15:10.840 --> 0:15:15.360 it's existential to humanity and it could destroy humanity. Not 0:15:15.480 --> 0:15:16.400 a small concern either. 0:15:16.440 --> 0:15:20.800 Yeah, because that's a certain thing, is it'd be in 0:15:20.840 --> 0:15:25.280 looking at water consumption, energy consumption footprint, there's also just 0:15:25.640 --> 0:15:27.920 it's fueled by a general apprehension. 0:15:28.160 --> 0:15:30.840 And it'd be like, so that it can take my job. 0:15:31.040 --> 0:15:33.400 Yeah, right, I got to give up all that, I 0:15:33.400 --> 0:15:35.560 got to give up clean water, I got to give 0:15:35.600 --> 0:15:37.960 up all this stuff. I got to accept this I 0:15:38.240 --> 0:15:41.360 sore into my community so that then this can take 0:15:41.440 --> 0:15:42.119 my friend's. 0:15:41.960 --> 0:15:47.040 Jobs, that's right, and my future, my children's future jobs. 0:15:47.200 --> 0:15:47.400 Yeah. 0:15:48.880 --> 0:15:50.600 And I don't think that the tech industry has done 0:15:50.600 --> 0:15:53.800 itself any favors by you know, carrying out a narrative 0:15:53.880 --> 0:15:57.960 that is you know, potentially quite exaggerated in that view, 0:15:58.000 --> 0:16:00.440 but you know, candidly, we all have a lot to 0:16:00.520 --> 0:16:05.400 learn as time unfolds. So so yeah, I mean you 0:16:05.480 --> 0:16:08.800 ask like, how do you separate or or figure out 0:16:08.920 --> 0:16:12.160 what is truth and what is what is crazy? And 0:16:14.240 --> 0:16:18.360 what what is rational to worry about? And I guess 0:16:18.400 --> 0:16:22.440 I I would start with what what is rational to 0:16:22.440 --> 0:16:26.400 worry about? It is very rational to worry about, you know, 0:16:27.440 --> 0:16:33.640 to fear change and uncertainty. And we are facing an incredible, 0:16:34.120 --> 0:16:39.000 uh turning point. We're in the early days of the 0:16:39.040 --> 0:16:42.400 Ai Revolution, and there's incredible fear and uncertainty, and that's rational. 0:16:43.520 --> 0:16:47.000 And I think it's rational to love your community and 0:16:47.080 --> 0:16:50.560 to love the character of your community. And you know, 0:16:51.320 --> 0:16:55.880 especially if you've you know, uh, even if there's been 0:16:55.920 --> 0:16:58.320 some economic decline, you've kind of made peace with that, 0:16:58.520 --> 0:17:00.360 or you you kind of love where you're from and 0:17:00.800 --> 0:17:02.360 you kind of enjoy a little bit of the piece 0:17:02.440 --> 0:17:05.399 that has come with some of the economic decline in 0:17:05.440 --> 0:17:10.000 your region and the natural beauty instead making you know, 0:17:11.000 --> 0:17:15.240 like coming to grips with it looking completely different. Is 0:17:16.080 --> 0:17:20.000 I think that's very rational. Like some of the scale 0:17:20.040 --> 0:17:23.800 of what I'm talking about in rural communities we are 0:17:23.840 --> 0:17:27.960 talking about in the end, a campus that I sometimes 0:17:28.040 --> 0:17:32.720 think of it as like, you know, five pentagons, like 0:17:32.920 --> 0:17:36.320 landing in the field across from your house. These are 0:17:36.640 --> 0:17:41.919 huge campuses, some of them. Yeah, like uh, well, uh, 0:17:42.600 --> 0:17:46.359 it's many more than Walmart super campuses. Like a Meta 0:17:46.400 --> 0:17:52.680 campus in Louisiana is five miles long and one mile wide, 0:17:52.720 --> 0:17:55.879 are you kidding? And so I've driven the perimeters of 0:17:56.080 --> 0:18:01.040 concrete not a concrete it's many buildings. So it's many 0:18:01.080 --> 0:18:06.520 buildings with land in between. And when they're finished, they 0:18:06.560 --> 0:18:09.760 will have some grass and they will have some parking lots, 0:18:09.760 --> 0:18:10.120 and they will. 0:18:10.400 --> 0:18:13.240 The footprint of that thing is five miles by a mile. Yeah, 0:18:13.359 --> 0:18:17.760 that particularly it's the footprint. It's it's the footprint of 0:18:17.800 --> 0:18:19.200 a large municipal airport. 0:18:19.960 --> 0:18:24.679 Yes, it's not dissimilar from an airport. And so it 0:18:24.840 --> 0:18:31.639 is rational to say, wait, this scale is truly it 0:18:31.720 --> 0:18:36.080 feels nuts to me, and and and and you know 0:18:36.119 --> 0:18:40.000 the truth is that the landscape has changed through many 0:18:40.080 --> 0:18:45.400 years of other industrial revolutions. Railroads have you know, crisscrossed 0:18:45.440 --> 0:18:47.840 and and you know, maybe maybe some thought at the 0:18:47.840 --> 0:18:50.160 time scarred the landscape and we got used to those 0:18:51.040 --> 0:18:51.840 we have area. 0:18:52.040 --> 0:18:56.640 Someone n like today was proposedly the railroad infrastructure as 0:18:56.680 --> 0:18:57.879 a new concept. 0:18:58.320 --> 0:18:59.240 You'd never get it done. 0:18:59.320 --> 0:19:00.400 You'd never get it on today. 0:19:00.600 --> 0:19:02.240 Like, hey, I got an idea I never run. 0:19:02.520 --> 0:19:04.879 You know, we're going to run these continuous lines all 0:19:04.880 --> 0:19:07.040 across the country and eminent domain everything. 0:19:07.560 --> 0:19:08.360 You'd never get it through. 0:19:08.440 --> 0:19:10.880 I'm going to put poles up with wires all over 0:19:10.920 --> 0:19:11.840 every single street. 0:19:12.480 --> 0:19:14.320 Polled wires and people would be like, no, what you're 0:19:14.320 --> 0:19:14.840 not doing that? 0:19:15.040 --> 0:19:17.520 Yeah, So I mean I think it's it's it's rational 0:19:17.600 --> 0:19:22.760 to you know, uh, and it's it's it's understandable too 0:19:22.920 --> 0:19:27.000 that people see technology and these they see data centers 0:19:27.040 --> 0:19:29.920 as see embodiment of how tech has changed our lives 0:19:30.240 --> 0:19:34.400 and you know, for the worse perhaps, and they sort 0:19:34.400 --> 0:19:36.320 of see it for the worse, and of course tech 0:19:36.320 --> 0:19:38.480 has also changed our lives for the better. And so 0:19:39.400 --> 0:19:42.560 I guess I would say what is not rational is 0:19:42.600 --> 0:19:46.560 to not look at and do this with some common 0:19:46.600 --> 0:19:49.480 sense and put context around it. So you know, one 0:19:49.480 --> 0:19:52.000 piece of context, right, like, well, look we now have 0:19:52.080 --> 0:19:55.439 airports that are these giant pieces of land that we 0:19:55.520 --> 0:19:59.680 all depend on. The truth is we do all depend 0:19:59.720 --> 0:20:03.760 on day centers every single day. Our lives are you know, 0:20:03.920 --> 0:20:07.320 as dependent on energy as they are in digital and 0:20:07.480 --> 0:20:14.200 now AI and energy are completely intertwined and they pull 0:20:14.200 --> 0:20:17.240 in all kinds of other resources and we're grappling with 0:20:17.320 --> 0:20:20.800 that and in real time in a hurry. The tech 0:20:20.800 --> 0:20:24.280 industry is in a big hurry. And that's another part 0:20:24.280 --> 0:20:30.240 of people's unease is how fast fast. 0:20:30.359 --> 0:20:32.960 Let's talk about some of the specifics, some of the 0:20:32.960 --> 0:20:40.399 specific impacts. Is we start with water, Okay, what amount 0:20:40.440 --> 0:20:43.640 of water? What amount of water are we talking about? 0:20:43.720 --> 0:20:46.520 I mean, let's look at it. You know, there's another 0:20:46.640 --> 0:20:48.600 question I got prior to that, because it just kind 0:20:48.600 --> 0:20:52.960 of sets this whole thing up. When we talk about 0:20:54.320 --> 0:21:00.680 what's going to happen, what is a comfortable timeline ten years? 0:21:00.680 --> 0:21:02.879 We have no idea, right so there's sort of like 0:21:02.960 --> 0:21:06.800 what's proposed right now, what's in process right now? So 0:21:06.880 --> 0:21:10.320 if I say to you how much water our data 0:21:10.320 --> 0:21:14.520 center is going to use? At what date are we 0:21:14.600 --> 0:21:15.120 looking at? 0:21:15.920 --> 0:21:16.000 Like? 0:21:16.080 --> 0:21:17.879 What date is it fair to look at? Is it 0:21:17.920 --> 0:21:19.920 like right now? Is it two years from now? 0:21:19.920 --> 0:21:22.000 People use twenty thirty as the. 0:21:22.520 --> 0:21:27.320 Guidepost, Okay, so when we're talking about it's just for 0:21:27.400 --> 0:21:31.640 people to reminder here we're in twenty twenty six. When 0:21:31.680 --> 0:21:35.920 we're talking about how much water, how much footprint, we're 0:21:35.960 --> 0:21:40.240 sort of looking at reasonable projections about four years from now. 0:21:40.840 --> 0:21:46.440 Is that cool? Okay? So when I read articles. 0:21:47.720 --> 0:21:52.000 Talking about data centers and resistance to data centers and things, 0:21:52.200 --> 0:21:55.160 one of the big issues that comes up is where 0:21:55.280 --> 0:21:57.720 is all this water coming from? And that winds up 0:21:57.760 --> 0:22:01.679 being like relevant. It's relevant to all Americans because we 0:22:01.720 --> 0:22:05.760 all rely on abundant clean water. It's relevant to agriculture 0:22:05.880 --> 0:22:09.440 obviously because food, and then in a sort of niche way, 0:22:10.040 --> 0:22:12.960 it's very relevant to like me as a fisherman and 0:22:13.000 --> 0:22:16.040 an outdoorsman, you know, like like the fact that our country, 0:22:16.119 --> 0:22:20.000 as wealthy as it is and as populous as it is, 0:22:20.000 --> 0:22:23.960 still has this like abundance of clean water full of 0:22:24.000 --> 0:22:27.639 fish relative to other places around the world. That's like 0:22:27.680 --> 0:22:30.400 a tremendous value. So when I hear something that might 0:22:30.480 --> 0:22:35.040 threaten that, I get anxious. You know, how do we 0:22:35.080 --> 0:22:36.320 look at the water issue here? 0:22:37.240 --> 0:22:41.000 Yeah, so I think so I just a couple of 0:22:41.040 --> 0:22:45.560 caveats that I'm not a water scientist, and and I'm 0:22:45.600 --> 0:22:48.879 also in the process as I work on this book 0:22:48.960 --> 0:22:52.240 of interrogating this topic from every angle that I can, 0:22:52.880 --> 0:22:57.280 and you know, striving to figure out the best version 0:22:57.320 --> 0:23:00.560 of the truth so that you know, we can all 0:23:00.600 --> 0:23:05.440 talk apples to apples. And so I've attempted recently did 0:23:05.440 --> 0:23:09.880 a column on water for the Information and have attempted 0:23:09.920 --> 0:23:13.320 to do that for the book as well, sort of 0:23:13.359 --> 0:23:19.280 saying what how much water, you know, does this industry 0:23:19.920 --> 0:23:24.760 really consume? And has that changed? So it has changed. 0:23:24.800 --> 0:23:29.359 Actually the data centers are getting bigger, but they're getting 0:23:29.400 --> 0:23:33.800 more water efficient. So let's put kind of water broadly 0:23:33.840 --> 0:23:37.560 speaking in perspective, so that this is this is kind 0:23:37.600 --> 0:23:41.000 of a rough exercise. And again, like you know, there 0:23:41.000 --> 0:23:45.879 are many resources that folks who have like helped me 0:23:45.880 --> 0:23:50.720 think through this are drawing from. Okay, but so we 0:23:50.760 --> 0:23:54.080 don't get alarmed by contextless numbers. You hear something like 0:23:55.040 --> 0:23:57.919 billions of gallons of water being used a year, and 0:23:57.960 --> 0:24:03.920 that sounds astonishingly large, sure, but in fact, big industries, 0:24:04.520 --> 0:24:08.440 including agriculture, are using trillions of gallons a year. 0:24:08.920 --> 0:24:10.800 So the other thing you hear, and is kind of glib, 0:24:11.200 --> 0:24:11.879 is you'll do it. 0:24:12.000 --> 0:24:15.840 You'll do a search, right, you do like a thing 0:24:15.880 --> 0:24:18.280 on chat GPT, and then people be like, well, there 0:24:18.320 --> 0:24:22.159 goes a gallon of water. Yeah, yeah, you hear it, 0:24:22.200 --> 0:24:24.440 but you don't really no one knows what I should say, 0:24:24.440 --> 0:24:24.680 No one. 0:24:24.760 --> 0:24:26.600 Most people don't really know what they're saying. 0:24:26.680 --> 0:24:28.840 Yeah, okay, so let's start big and we kind of 0:24:28.920 --> 0:24:31.439 telescope you know, out and then go go in a 0:24:31.440 --> 0:24:35.200 little bit. So there's like, you know, if you maybe 0:24:35.200 --> 0:24:36.760 you can keep a number in your head of like 0:24:37.200 --> 0:24:40.879 fresh water consumption in the US and the like thirty 0:24:40.960 --> 0:24:46.520 trillion gallons a year range, and. 0:24:45.520 --> 0:24:47.159 Well that's what that's what we're using. 0:24:47.480 --> 0:24:54.440 Yeah, and you know, and and thirty gallons thirty trillion 0:24:54.480 --> 0:24:59.880