WEBVTT - Inside Amazon’s Push for Sustainability 0:00:00.030 --> 0:00:04.570 We're very stubborn on our long-term vision to get to 0:00:04.910 --> 0:00:09.340 net zero carbon by 2040. That is a very difficult goal. 0:00:10.220 --> 0:00:14.320 It's a really tough challenge. We set that goal in 2019. 0:00:14.320 --> 0:00:16.000 And for a company of our size and our scale 0:00:16.040 --> 0:00:18.140 to achieve that is going to be really hard. 0:00:22.850 --> 0:00:25.430 I'm Laurie Siegel, and you're listening to Mostly Human, a 0:00:25.470 --> 0:00:31.240 tech podcast through a human lens. Cara, I'm so excited 0:00:31.360 --> 0:00:32.560 to have you on Mostly Human. 0:00:32.580 --> 0:00:34.880 Yeah, well, thanks for having me. I'm really excited to 0:00:34.900 --> 0:00:36.080 have this conversation with you. 0:00:36.360 --> 0:00:39.820 You are sitting at such an important seat at the 0:00:39.900 --> 0:00:43.450 table at such an important moment. You're the chief sustainability 0:00:43.479 --> 0:00:46.269 officer at Amazon, meaning you are one of the people 0:00:46.310 --> 0:00:49.110 in charge of helping one of the biggest companies. be sustainable, 0:00:49.210 --> 0:00:52.089 but also it has implications for all of us. You know, 0:00:52.130 --> 0:00:53.709 before we get into it fully, I want to say 0:00:53.750 --> 0:00:56.330 that we're partnering with Amazon for this episode because I 0:00:56.630 --> 0:00:59.990 really believe that sustainability and I have, I was just 0:01:00.010 --> 0:01:01.470 saying to you before this started, I have a 20 0:01:01.470 --> 0:01:04.510 month old, like I feel like sustainability is one of 0:01:04.530 --> 0:01:08.250 the most important lenses we can look at the future 0:01:08.380 --> 0:01:10.880 through for my son growing up, for our children growing 0:01:10.920 --> 0:01:13.580 up in the world that we're leaving behind. And so 0:01:13.620 --> 0:01:17.220 it feels important for folks to hear from you at 0:01:17.260 --> 0:01:19.940 this moment and in this role because of the power 0:01:20.000 --> 0:01:22.900 and the influence of it. And I'd love to start 0:01:23.080 --> 0:01:27.259 with talking about the contradictions of this moment. We've built 0:01:27.420 --> 0:01:30.440 a world which is extraordinary. In the tap of a button, 0:01:30.940 --> 0:01:34.730 anything can arrive at our doorsteps, right? But we also 0:01:34.790 --> 0:01:38.030 ask the question, what are we asking the planet to absorb? 0:01:38.230 --> 0:01:41.830 AI is an incredible solution. It's also a problem in 0:01:41.870 --> 0:01:45.190 many ways. It could fight against climate change, but also 0:01:45.530 --> 0:01:49.330 powering AI contributes to it. And we hear a lot 0:01:49.350 --> 0:01:51.370 of solutions from big tech. And sometimes it feels like 0:01:51.390 --> 0:01:52.990 we're just running out of time when you look at 0:01:53.050 --> 0:01:56.790 some of the headlines. I know that this isn't something 0:01:57.070 --> 0:02:00.470 that you just walked into. Even when you walked into 0:02:00.500 --> 0:02:04.380 the doors of Amazon back in 2014, you have a 0:02:04.500 --> 0:02:09.980 long history before Amazon, nonprofit, government, technology. So you've been 0:02:10.040 --> 0:02:13.750 thinking about this for a very long time. what do 0:02:13.770 --> 0:02:16.710 you make of this very specific moment that we're in, 0:02:16.990 --> 0:02:17.649 in 2026? 0:02:17.650 --> 0:02:22.750 Yeah, I think one of the benefits that I hopefully 0:02:22.770 --> 0:02:25.130 bring to this role, but I think a lot of 0:02:25.180 --> 0:02:27.579 people I work with do as well, is that kind 0:02:27.660 --> 0:02:31.970 of multi-sector thinking. So I had worked in many nonprofits. 0:02:32.190 --> 0:02:34.470 I've worked for government. I've worked for a senator. I've 0:02:34.490 --> 0:02:38.310 worked for a mayor. And I'd worked with corporations before 0:02:38.820 --> 0:02:42.560 in many different roles, thinking about sustainability for different types 0:02:42.600 --> 0:02:45.490 of industries. And so I think you have to be 0:02:45.530 --> 0:02:49.290 a systems thinker to think about sustainability, because you have 0:02:49.330 --> 0:02:51.510 to think about all of the different ways that we 0:02:51.570 --> 0:02:55.230 impact the planet, the different technologies that we're using, the 0:02:55.250 --> 0:02:57.530 different ways that you might construct a product and what 0:02:57.570 --> 0:03:01.130 that means. You have to think very holistically about so 0:03:01.150 --> 0:03:03.120 many different impacts. A lot of what we do is 0:03:03.160 --> 0:03:05.700 think about life cycle thinking, you know, as an example, 0:03:05.720 --> 0:03:07.140 and we can talk a little bit about that and 0:03:07.160 --> 0:03:10.400 what that means. But one of the things that we 0:03:10.440 --> 0:03:13.799 really try to do is to bring that Systems thinking 0:03:13.900 --> 0:03:16.940 to all the ways in which we can bring all 0:03:16.960 --> 0:03:19.630 of the skills and the advantages that we have at 0:03:19.710 --> 0:03:22.889 Amazon to how we can bring that technological innovation to 0:03:23.330 --> 0:03:27.380 the problems at hand, as you pointed out. some of 0:03:27.440 --> 0:03:30.970 the innovations that we're all driving, they have benefits and 0:03:30.990 --> 0:03:34.670 they have opportunities. AI can bring incredible ways to advance 0:03:34.710 --> 0:03:38.830 sustainability and AI compute also brings a big demand for energy, 0:03:39.170 --> 0:03:41.650 which can bring a lot of issues and, you know. 0:03:41.720 --> 0:03:42.680 Demands on the grid. 0:03:43.120 --> 0:03:45.680 So, you know, there's pros and cons of a lot 0:03:45.700 --> 0:03:47.740 of the things that, you know, we're seeing develop in 0:03:47.760 --> 0:03:51.050 the world today. Um, And we want to look at 0:03:51.150 --> 0:03:53.770 both of those things and we want to operate responsibly. 0:03:53.930 --> 0:03:56.960 And so that's really the joy of working at a 0:03:57.000 --> 0:03:59.880 place like Amazon is that we have that in our DNA, 0:03:59.940 --> 0:04:04.700 I think sustainability. And by that, I really mean environmental responsibility, 0:04:05.200 --> 0:04:09.150 attention to human rights, thinking about things around people and planet, 0:04:09.810 --> 0:04:13.830 that that is how we fundamentally operate because we think 0:04:13.910 --> 0:04:15.510 about the communities where we live. 0:04:15.410 --> 0:04:16.190 And work and operate. 0:04:16.230 --> 0:04:18.630 So we want to bring that kind of systems thinking 0:04:18.650 --> 0:04:21.539 to everything that we're doing and think about it at 0:04:21.580 --> 0:04:23.500 its core. How do we serve our customer? How do 0:04:23.540 --> 0:04:26.280 we serve our community? And how do we design to 0:04:26.380 --> 0:04:30.100 operate responsibly? So that requires when we think about all 0:04:30.140 --> 0:04:33.250 of those different industries and products and services and what 0:04:33.310 --> 0:04:36.589 they create in the world, that requires going to first 0:04:36.630 --> 0:04:39.770 principles and thinking about how do we design the products 0:04:39.790 --> 0:04:42.229 that we make? What are the materials that we use? 0:04:42.250 --> 0:04:45.190 What are the impact those materials are gonna have? What's 0:04:45.210 --> 0:04:48.400 the disposal end of life going to look like? Where 0:04:48.420 --> 0:04:51.589 does all of that stuff go? How do we design 0:04:51.650 --> 0:04:55.049 for less waste? How do we design for our customers 0:04:55.089 --> 0:04:58.240 at the end of the day? It really requires very 0:04:58.320 --> 0:05:01.900 long-term thinking and, you know, not quarter to quarter. 0:05:03.000 --> 0:05:06.440 It's interesting because I saw something you wrote in a 0:05:06.500 --> 0:05:09.100 public letter. You said, while the speed and scale of 0:05:09.140 --> 0:05:11.430 adoption is unique and the change is happening faster and 0:05:11.450 --> 0:05:14.800 more broadly than anything we've encountered in our lifetimes, The 0:05:14.839 --> 0:05:17.500 need to stay stubborn on our vision and flexible on 0:05:17.520 --> 0:05:21.409 the details is familiar territory. So what, Kara, are you 0:05:21.470 --> 0:05:23.830 staying stubborn on right now as we look at this? 0:05:23.970 --> 0:05:26.290 And what details are you flexible on? 0:05:26.550 --> 0:05:31.650 Yeah, we're very stubborn on our long-term vision to get 0:05:31.750 --> 0:05:35.700 to net zero carbon by 2040. That is a very 0:05:35.900 --> 0:05:39.890 difficult goal. It's a really tough challenge. We set that 0:05:39.930 --> 0:05:42.950 goal in 2019. And for a company of our size 0:05:42.980 --> 0:05:44.560 and our scale to achieve that is. 0:05:44.520 --> 0:05:47.740 Going to be really hard. We don't know all of 0:05:47.800 --> 0:05:49.260 the ways that we're going to get there. 0:05:49.279 --> 0:05:52.830 And every year that we look at how we're going 0:05:52.850 --> 0:05:56.290 to do that, the details change. We set again, like 0:05:56.330 --> 0:05:58.290 that's something where we set it in 2019 for a 0:05:58.290 --> 0:06:00.250 company of our size and scale and all of the 0:06:00.290 --> 0:06:02.140 different ways that we're going to change between when we 0:06:02.180 --> 0:06:04.780 set it in 2019 and going to 2040 to set 0:06:04.820 --> 0:06:06.420 that kind of long term vision. 0:06:07.640 --> 0:06:08.540 Really challenging. 0:06:09.220 --> 0:06:11.990 But unless you set that kind of long term vision, 0:06:12.260 --> 0:06:15.310 you can't get everybody rowing in the same direction. And 0:06:15.350 --> 0:06:19.570 you can't also make, you know, decisions along the way. 0:06:20.140 --> 0:06:24.180 And time out those decisions and change big things like 0:06:24.420 --> 0:06:30.339 electrifying your entire fleet or moving decisions around ocean freight, 0:06:30.420 --> 0:06:32.500 which is going to take a long time to change 0:06:32.870 --> 0:06:36.229 ports and demand signals for ocean freight. I mean, think 0:06:36.270 --> 0:06:38.570 about that industry and how hard that would be to change. 0:06:39.250 --> 0:06:42.550 Or things like how to build differently. We think about 0:06:42.630 --> 0:06:47.210 data centers. It's not just the energy demands and procuring 0:06:47.520 --> 0:06:54.680 carbon-free energy, but Literally, cement, steel, concrete, those industries need 0:06:54.720 --> 0:06:58.039 to change and they need to provide better materials, which 0:06:58.080 --> 0:06:59.720 means we need to demand them and then we need 0:06:59.760 --> 0:07:02.980 to build with those materials. We need to invent new 0:07:03.020 --> 0:07:05.359 ways of, you know, we have our own chips and 0:07:05.400 --> 0:07:09.500 we run silicon and we need to invent and keep 0:07:09.660 --> 0:07:13.460 reinventing ways to get more efficient with that. So all 0:07:13.480 --> 0:07:17.110 of these things need to change and have been. So 0:07:17.230 --> 0:07:20.050 I can talk about, again, the specific ways that we're 0:07:20.070 --> 0:07:24.270 building better and getting more efficient, creating new products. But 0:07:24.450 --> 0:07:29.440 I think the path that we're on, all of those 0:07:29.500 --> 0:07:33.120 details about how we're building, how we're operating, the materials 0:07:33.160 --> 0:07:35.710 that we're using, whether we're building a fulfillment center, a 0:07:35.730 --> 0:07:39.470 data center, or a Kindle, are changing along the way. 0:07:39.490 --> 0:07:42.670 And we're learning as we go. But that long term 0:07:42.730 --> 0:07:43.730 vision stays the same. 0:07:43.870 --> 0:07:45.490 And could you tell me a little bit about the 0:07:45.530 --> 0:07:48.290 learning curve? So, you know, you talk about Amazon's goals 0:07:48.350 --> 0:07:51.590 to reach net zero global emissions by 2040, which, as 0:07:51.610 --> 0:07:56.030 you say, is extraordinary, ambitious, right? Every year you put 0:07:56.070 --> 0:07:58.570 out a sustainability report with updates on how the company's 0:07:58.620 --> 0:08:02.520 meeting those goals. Talk about the learning curve. on sustainability, 0:08:02.900 --> 0:08:04.200 on the sustainability goal? 0:08:04.360 --> 0:08:06.700 Yes, we have to think about setting the goal and 0:08:06.720 --> 0:08:09.119 then working backwards from that, which we do at Amazon 0:08:09.160 --> 0:08:09.440 a lot. 0:08:09.480 --> 0:08:11.500 We work backwards from our customer. 0:08:11.560 --> 0:08:13.970 We work backwards from a goal and think about how 0:08:13.990 --> 0:08:16.570 do we operationalize that and mechanize it. We're really good 0:08:16.590 --> 0:08:19.150 at building mechanisms at Amazon. We're really good at thinking 0:08:19.190 --> 0:08:23.010 about putting action, you know, action around those kind of goals. 0:08:23.490 --> 0:08:24.870 And so what we do is we sit down and 0:08:24.910 --> 0:08:27.650 we work with, as you said, each business unit. So 0:08:27.680 --> 0:08:29.860 I'm going to sit down with our devices unit or 0:08:29.920 --> 0:08:35.110 Amazon Web Services unit or studios, or our retail unit 0:08:35.150 --> 0:08:37.150 and say, okay, what's your business. 0:08:36.910 --> 0:08:37.650 Plan for the year? 0:08:37.670 --> 0:08:41.210 And what kind of goals could you achieve? And what's 0:08:41.270 --> 0:08:44.610 the biggest carbon lever that you could pull? So it 0:08:44.670 --> 0:08:48.550 might be for our retail fulfillment logistics business, might be 0:08:48.590 --> 0:08:51.349 fleet electrification. It might be saying, okay, we've got a 0:08:51.390 --> 0:08:54.110 lot of delivery vehicles. So let's think about how we're 0:08:54.130 --> 0:08:57.650 gonna procure more electric delivery vans. It might be better 0:08:57.670 --> 0:09:00.530 building materials because buildings have a lot of You know, 0:09:00.570 --> 0:09:02.550 carbon impact. So we want to say we're going to 0:09:02.570 --> 0:09:05.000 build with mass timber or we're going to procure more 0:09:05.020 --> 0:09:09.260 low carbon concrete. It might be swapping out, you know, 0:09:09.460 --> 0:09:11.120 as we did in the early days. There's a lot 0:09:11.140 --> 0:09:14.560 of winds of swapping out lighting. That's an easy one. Right. 0:09:14.580 --> 0:09:15.680 Or HVAC systems. 0:09:16.410 --> 0:09:18.250 We did a lot of putting in a building management 0:09:18.270 --> 0:09:21.750 systems because we can just manage heating and we can manage. 0:09:21.630 --> 0:09:22.610 Electricity that way. 0:09:22.790 --> 0:09:25.709 We procure a lot of carbon-free energy, wind, solar, and 0:09:25.730 --> 0:09:28.689 now increasingly nuclear deals. But then we think about Kindles. 0:09:28.830 --> 0:09:31.730 There's a roadmap for making any product that we're going 0:09:31.750 --> 0:09:35.310 to manufacture. And then that product roadmap has to back 0:09:35.410 --> 0:09:37.870 out a number of months and think about, okay, how 0:09:37.910 --> 0:09:40.050 do we begin to swap out not only the materials 0:09:40.070 --> 0:09:43.090 that go into the Kindle, but also the packaging materials 0:09:43.330 --> 0:09:45.960 and make sure that's a great customer experience. but that 0:09:45.980 --> 0:09:48.400 we can make it more sustainable. For studios, you think 0:09:48.440 --> 0:09:51.040 about all the things that go into our production. So 0:09:51.120 --> 0:09:55.450 how do we have a less wasteful set? How do 0:09:55.470 --> 0:09:58.590 we think about everything from craft services that might get 0:09:58.630 --> 0:10:01.140 set up to actually building a set? And then along 0:10:01.160 --> 0:10:03.839 the way, when we started out in 2019, we didn't 0:10:03.880 --> 0:10:04.560 have Whole Foods. 0:10:05.360 --> 0:10:06.120 Now we have 500 Whole. 0:10:05.960 --> 0:10:11.040 Foods physical grocery stores. That's a new challenge. So how 0:10:11.059 --> 0:10:12.580 do you think about all of a sudden we had 0:10:12.660 --> 0:10:15.260 physical retail? How do you think about all those stores 0:10:15.420 --> 0:10:19.120 and everything that goes into agricultural commodity supply chains and 0:10:19.160 --> 0:10:22.060 all the impacts that that brings. Luckily, they were very 0:10:22.100 --> 0:10:24.400 sustainably minded, and so they brought a lot of knowledge 0:10:24.460 --> 0:10:27.590 as well. But any new business that we get into, 0:10:27.640 --> 0:10:30.130 we have to think about what they bring, what are 0:10:30.170 --> 0:10:33.770 the impacts, and then how can we really learn about 0:10:33.830 --> 0:10:35.390 the sustainability opportunities there. 0:10:35.410 --> 0:10:37.060 And can you walk us through some of the really 0:10:37.100 --> 0:10:39.300 specific examples? I mean, I know a passion of yours 0:10:39.340 --> 0:10:42.000 is water and the role it plays in economic development. 0:10:42.059 --> 0:10:45.520 You're on the board of water.org. I think I saw 0:10:45.570 --> 0:10:46.990 you do the hot week. 0:10:47.150 --> 0:10:47.750 I did. 0:10:48.710 --> 0:10:51.689 You have to really love water to go and try 0:10:51.730 --> 0:10:53.770 to promote it by eating like hot wings. 0:10:53.950 --> 0:10:56.530 Yeah. I don't like spicy food either. 0:10:56.640 --> 0:11:01.380 Yeah. And that's for reference, like Matt Damon's organization. And 0:11:01.420 --> 0:11:03.480 so talk to me a little bit about the passion 0:11:03.500 --> 0:11:05.700 for water, the role it plays in economic development. And 0:11:05.730 --> 0:11:07.429 if you could get into specifically some of the work 0:11:07.530 --> 0:11:10.020 that Amazon does. has done here. 0:11:10.300 --> 0:11:12.640 That's a great example of an issue that I think 0:11:12.760 --> 0:11:15.840 wasn't actually on the corporate radar for a long time 0:11:16.220 --> 0:11:18.719 in the sense that we take water, especially in the 0:11:18.740 --> 0:11:21.120 United States, we take water for granted. It's not something 0:11:21.160 --> 0:11:23.910 that people think about, but one in four people around 0:11:23.950 --> 0:11:26.329 the world don't have access to safe water. And that 0:11:26.410 --> 0:11:30.069 is an astounding statistic when you think about what water 0:11:30.110 --> 0:11:31.970 does for us every day in our lives and what 0:11:32.010 --> 0:11:34.880 we need water for in terms of, you know, having 0:11:35.160 --> 0:11:38.480 water just to keep ourselves healthy and, you know, for sanitation, 0:11:38.580 --> 0:11:42.480 for our own personal health, and then really to be 0:11:42.520 --> 0:11:44.110 part of operating our lives. 0:11:44.370 --> 0:11:46.089 And to operate a business, you need water. 0:11:46.250 --> 0:11:52.689 So water has increasingly become a business issue. And we 0:11:52.730 --> 0:11:55.670 were talking about this at Amazon and talking about it 0:11:55.710 --> 0:11:58.050 in a number of different ways. You need water to 0:11:58.090 --> 0:12:01.670 make food product. You need water to run buildings. So 0:12:02.070 --> 0:12:05.410 as water was rising on the corporate agenda, we were 0:12:05.510 --> 0:12:07.790 looking at different ways that we could also think about 0:12:07.830 --> 0:12:12.030 water in a long-term way. And we've done a number 0:12:12.190 --> 0:12:17.800 of different things to invest in water and protect water. 0:12:17.840 --> 0:12:20.000 For the communities where we operate. 0:12:20.360 --> 0:12:24.140 One is that for our data center business, and that's 0:12:24.179 --> 0:12:26.839 where water has been a very hot issue. There's a 0:12:26.940 --> 0:12:30.520 lot of misunderstanding around water, I think, in relationship to 0:12:30.559 --> 0:12:35.630 data centers. There's a lot of, I think, ways in 0:12:35.690 --> 0:12:40.010 which You can operate very responsibly with regards to water. 0:12:40.090 --> 0:12:43.050 So we set a goal to be water positive in 0:12:43.090 --> 0:12:45.569 our Amazon Web Services business. And what that means is 0:12:45.690 --> 0:12:47.700 anywhere where that we have a data center operation, we're 0:12:47.720 --> 0:12:50.800 going to return more water to that community where we 0:12:50.860 --> 0:12:53.720 operate a data center than we utilize ourselves. And we 0:12:53.760 --> 0:12:55.559 are going to be water positive by 2030. We're already 75% 0:12:55.559 --> 0:13:00.580 of the way there. We have over 70 water replenishment 0:13:00.600 --> 0:13:04.210 projects around the world. But we also, 90% of the 0:13:04.250 --> 0:13:06.610 time that we're operating our data centers, we don't use 0:13:06.770 --> 0:13:07.689 any water at all. 0:13:08.790 --> 0:13:11.619 Explain that. I mean, because I look at even in 0:13:11.660 --> 0:13:15.590 doing research for this, look and see a stat that says, 0:13:16.150 --> 0:13:18.010 from the Lincoln Institute of Land Policy, it says a 0:13:18.070 --> 0:13:20.970 mid-sized data center consumes as much water as a small town, 0:13:21.010 --> 0:13:23.589 while larger ones require up to 5 million gallons of 0:13:23.670 --> 0:13:27.290 water every day, as much as a city of 50,000 people. 0:13:27.370 --> 0:13:30.230 So when you hear stats like that, I mean- There's. 0:13:30.050 --> 0:13:32.800 A lot of misinformation in that. There's a wide range 0:13:32.840 --> 0:13:35.120 of how you can run and operate a data center. 0:13:35.600 --> 0:13:38.120 We build, I'll talk about ours only. I can't speak 0:13:38.140 --> 0:13:40.900 to everyone's data center and how they build, but you can, 0:13:41.200 --> 0:13:44.200 if you choose so, You can build really efficiently. The 0:13:44.240 --> 0:13:47.360 technology is available to you. So how we choose to 0:13:47.400 --> 0:13:51.520 build is most of the time we use air cooling. 0:13:51.620 --> 0:13:54.410 So you can bring in air and use, water's used 0:13:54.429 --> 0:13:56.870 to cool. So let's start with that. Why do you 0:13:56.910 --> 0:13:58.490 need water in a data center, right? 0:13:58.650 --> 0:14:00.190 We can, you know, and it might even be worth 0:14:00.230 --> 0:14:03.210 taking a step back to talk about data centers in general, 0:14:03.230 --> 0:14:04.550 because I do want to, and I do want to 0:14:04.570 --> 0:14:06.870 get more into some of the specifics that Amazon's doing 0:14:06.890 --> 0:14:08.710 around disaster relief, a lot of this stuff. But let's, 0:14:08.830 --> 0:14:10.770 I'd love to do, let's do the data center stuff 0:14:10.830 --> 0:14:13.420 because it is the type of thing Even when I 0:14:13.460 --> 0:14:16.660 tell people I'm interviewing the chief sustainability at Amazon, everyone's like, 0:14:16.920 --> 0:14:18.480 I hope you're going to ask about data centers. 0:14:18.540 --> 0:14:18.800 Of course. 0:14:33.560 --> 0:14:35.780 It's happening for some people, but not all. And it's 0:14:35.840 --> 0:14:38.420 happening at the cost of, could it happen at the 0:14:38.440 --> 0:14:41.020 cost of our jobs and our planet and whatnot? And 0:14:41.040 --> 0:14:44.240 it seems that it has been a physical, the physical 0:14:44.280 --> 0:14:48.180 manifestation of this anger and frustration has come, I think, 0:14:48.520 --> 0:15:04.620 in the form of these data centers popping up. Could 0:15:04.920 --> 0:15:08.200 we do, if you don't mind, and go with me 0:15:08.260 --> 0:15:08.680 for this. 0:15:08.710 --> 0:15:08.880 Sure. 0:15:09.280 --> 0:15:11.730 Could you, you know, do a little bit of a 0:15:11.830 --> 0:15:16.590 data center for dummies? Like, break down what they are, 0:15:16.810 --> 0:15:19.810 why they are needed, and how this has changed. Because 0:15:19.830 --> 0:15:22.380 AWS has been a leader in the space for a 0:15:22.390 --> 0:15:27.100 long time. Data centers are vital for cloud success. Now 0:15:27.140 --> 0:15:30.320 they're divisive. But explain, you know, to some degree. 0:15:30.660 --> 0:15:31.500 Yeah, we've been. 0:15:31.380 --> 0:15:35.400 Building data centers for, I mean, for a very long time. 0:15:35.460 --> 0:15:37.300 And help people understand why. Yeah. 0:15:37.740 --> 0:15:43.520 So data centers are buildings that hold server racks of servers. 0:15:43.600 --> 0:15:46.880 If you go in and tour a data center, what 0:15:46.900 --> 0:15:51.040 you'll see is basically I have been in one long 0:15:51.120 --> 0:15:56.620 hallways with lots of racks of servers and wires. And. 0:15:58.570 --> 0:16:01.130 Really what's happening in there is there's a lot of 0:16:01.550 --> 0:16:07.770 data being exchanged on those servers. And we need data 0:16:07.820 --> 0:16:11.520 centers for... everything that we're doing in our everyday lives. 0:16:11.640 --> 0:16:14.560 So if you are ordering an Uber, you're shopping for 0:16:14.580 --> 0:16:18.740 your groceries online, you're checking your kid's soccer schedule, a 0:16:18.820 --> 0:16:21.580 lot of the activities that we're doing in our everyday lives, 0:16:21.900 --> 0:16:24.900 I mean, you have a data plan on your phone, 0:16:25.280 --> 0:16:28.780 you're consuming data, we need data centers, right? These are, 0:16:29.360 --> 0:16:32.590 I think a lot of times we think about, there's 0:16:32.620 --> 0:16:35.310 a bit of a mystery about what happens, I think, 0:16:35.330 --> 0:16:38.950 sometimes in data centers. They are to process data. So 0:16:38.970 --> 0:16:41.970 to process data for what happens with us as individuals 0:16:42.030 --> 0:16:45.630 and to process data for companies that are processing large. 0:16:45.410 --> 0:16:46.630 Amounts of data, right? 0:16:47.050 --> 0:16:50.750 So if you're using a streaming service, if you're watching 0:16:50.790 --> 0:16:53.920 a movie, that's data that's going through and being processed 0:16:54.400 --> 0:16:56.960 in a data center. They have been in existence for 0:16:56.980 --> 0:17:00.540 a while. What you're seeing now with AI is because 0:17:00.580 --> 0:17:04.880 of the compute power needed for AI is more intense, right? 0:17:05.350 --> 0:17:08.210 And so more data centers are needed because more compute 0:17:08.230 --> 0:17:11.609 power is needed because AI is more intensive in the 0:17:11.650 --> 0:17:14.260 compute power needed. More energy is then needed to feed that. 0:17:14.300 --> 0:17:15.740 So more data centers are needed. 0:17:15.840 --> 0:17:18.100 And we have almost like this compute race in Silicon 0:17:18.140 --> 0:17:21.580 Valley of, you know, staying ahead, staying ahead. You know, 0:17:21.760 --> 0:17:25.460 all these companies, there's going to be, there could be 0:17:25.609 --> 0:17:27.350 IPOs in the next couple of years. And a lot 0:17:27.390 --> 0:17:28.930 of this depends on compute. 0:17:29.250 --> 0:17:29.810 On compute. 0:17:30.030 --> 0:17:34.470 And that can, what it should bring is, is responsible 0:17:34.520 --> 0:17:38.700 building and an economic boom for the United States. But 0:17:38.800 --> 0:17:44.520 I understand why communities are concerned about what it brings 0:17:44.619 --> 0:17:48.080 for them. And so one of the things that we 0:17:48.140 --> 0:17:53.260 have said at Amazon is that we really want to 0:17:53.300 --> 0:17:57.920 build and have been building responsibly. Now, I can't speak 0:17:57.960 --> 0:18:00.450 for every company that builds a data center, but this 0:18:00.530 --> 0:18:02.929 is where I think it is important for communities to 0:18:03.150 --> 0:18:05.230 ask very good questions. 0:18:05.310 --> 0:18:05.770 Like what? 0:18:06.010 --> 0:18:08.250 About how a data center is being built and what 0:18:08.290 --> 0:18:11.889 they get in return for that investment in having a 0:18:11.930 --> 0:18:15.050 data center in their community. So you should ask about, 0:18:15.900 --> 0:18:18.940 how that data center is being built. What kinds of 0:18:18.960 --> 0:18:24.020 materials are being used? What kinds of energy affordability questions 0:18:24.280 --> 0:18:27.580 need to be answered? What kind of water use questions 0:18:27.800 --> 0:18:31.820 as we were getting into, right? What are the sustainability 0:18:31.960 --> 0:18:34.920 aspects of how a company is going to operate? Not 0:18:34.940 --> 0:18:38.699 all data centers are the same. So we build very 0:18:38.770 --> 0:18:43.030 efficiently with sustainable materials. We have a commitment to carbon-free 0:18:43.109 --> 0:18:46.750 energy and a commitment to water positivity. which means we 0:18:46.790 --> 0:18:50.930 are going to build incredibly efficiently. We are 90% of 0:18:50.950 --> 0:18:53.520 the time not using any water at all in our 0:18:53.540 --> 0:18:58.100 data centers. And we have a commitment, again, to carbon-free energy. 0:19:00.430 --> 0:19:04.110 That means we're building the most sustainably, right? And so 0:19:04.250 --> 0:19:05.990 I can tell you about, if you wanna go into 0:19:06.010 --> 0:19:11.230 the details, our power use efficiency metric, our PUE is 1.14, 0:19:11.230 --> 0:19:15.399 which is very, very efficient in the way that we 0:19:15.500 --> 0:19:19.619 use power in our data centers. We build with, again, 0:19:20.660 --> 0:19:24.420 better materials. So we're looking at low carbon concrete, we're 0:19:24.440 --> 0:19:28.240 looking at low carbon steel. We look at, our chips 0:19:28.460 --> 0:19:32.719 are very energy efficient. We are not using water to cool, 0:19:33.000 --> 0:19:36.850 certainly not in any water-constrained areas. And data centers in 0:19:36.970 --> 0:19:41.510 general are using less water than a lot of industries, 0:19:41.650 --> 0:19:45.810 less water than almond farming, less water than golf courses. 0:19:46.210 --> 0:19:49.449 So I think there's a lot of attention on them, 0:19:49.470 --> 0:19:52.320 and I understand. why, because all of the AI and 0:19:52.340 --> 0:19:54.600 the compute and also the concerns about AI that are 0:19:54.619 --> 0:19:58.460 happening right now. But we have to also look at 0:19:58.520 --> 0:20:03.470 facts and also demand efficient building. And then I think 0:20:03.490 --> 0:20:05.380 there's the other side of it, which is also what 0:20:05.440 --> 0:20:09.060 is the economic benefit that's happening? So we've started to 0:20:09.119 --> 0:20:12.460 study also like you should have economic benefit for some 0:20:12.500 --> 0:20:15.320 of these investments. We've announced a series of benefits that 0:20:15.340 --> 0:20:17.699 we want to bring to data center communities. So we 0:20:17.740 --> 0:20:20.300 announced a program called Built Together, which means for our 0:20:20.340 --> 0:20:24.700 data center communities in the United States, we're offering free 0:20:24.740 --> 0:20:29.590 community college to residents of data center communities. We're offering 0:20:30.190 --> 0:20:32.850 modular training centers, which are going to be upskilling programs 0:20:32.869 --> 0:20:35.690 so people can participate in getting the skills needed to 0:20:35.730 --> 0:20:39.970 participate in the AI economy. We're offering energy affordability programs. 0:20:39.990 --> 0:20:42.760 So if you want to retrofit your home with heat 0:20:42.780 --> 0:20:45.320 pumps or, you know, energy affordability so you can keep 0:20:45.359 --> 0:20:48.820 your bills low, we've signed the Ratepayer Protection Act, which 0:20:48.859 --> 0:20:51.780 means we are committed to not having, you know, rates raised. 0:20:52.260 --> 0:20:56.100 We're also saying we think communities know best about what 0:20:56.140 --> 0:20:59.510 their community needs. And so we're partnering to give community 0:20:59.550 --> 0:21:02.350 grants to communities directly, give them money so that they 0:21:02.390 --> 0:21:05.130 can spend it on if they need roads or schools 0:21:05.150 --> 0:21:07.530 or whatever. They know best what they need. So I 0:21:07.570 --> 0:21:11.170 think all of those economic benefits, you should demand that 0:21:11.210 --> 0:21:14.480 for your community alongside of this. Like there should be 0:21:14.520 --> 0:21:16.580 a whole exchange that happens. 0:21:17.900 --> 0:21:20.199 How do you do that? It's so interesting because it's, 0:21:20.600 --> 0:21:22.520