Edge Computing: A Greener, Better Data Solution With Tom Frazier

We think that the opportunity to address sustainability and social impact is edge computing, which are smaller facilities that can go closer to where the source of the data is. So if you think about it, data is generated from somewhere and it's consumed from somewhere, and the shorter you can make that distance, the more efficient you can be across all of these concepts. And so by making facilities smaller, we can put them closer to the 5G tower, for example. So you get faster speeds, or they're closer to a smart city which doesn't want their data to leave their local area. And so all of our facilities are smaller in size. They're modular, so you can make one hold 10 racks worth of servers and you can make another one that holds 100 racks worth of servers. But we're not designing any of these to be the size of football fields.

Welcome to Care More, Be Better, a podcast for people like you who care about the social impact of conscious companies and everyday heroes. Hear inspiring stories from those who put people and planet before profit and personal gain. You'll learn how you can make a difference, vote with your dollars, and get involved today. Here's your host, Karina Belizzi. Hello, fellow do-gooders and friends. I'm your host, Karina Belizzi. Over the past few weeks, we really dove deep into subjects around building the future that we all want to live in. From building spaces and cities that we can really get around in easily, live, work, and play within, to the measures that we'll need to take to create resilient localized food systems. But living in the modern world will also need to make technology more accessible and greener. Data in all of its forms should be easy to use, fast, affordable, and leveraged for good. Joining me for this conversation is someone whose many talents range from competitive swing dancing to leading world-class data centers for Fortune 10 companies. Tom Frazier is the co-founder and CEO of Redivider, a sustainable and responsible edge computing company. So for those of you who don't understand what edge computing is, we'll get into that today too. Tom is devoted to spearheading innovation in this digital economy while leading with a commitment to people, planet, and profits. Tom Frazier, welcome to the show.

Thanks for having me. Very excited to be here.

Now, I think I can trust a competitive swing dancer to keep this conversation, let's say, interesting.

I hope so. I hope so.

I'll just ask that you don't bore me to death the way my husband does with details when it comes to things like eVTOL, electric vertical takeoff and landing crafts, because that's the space that he's in, so tech and aerospace. But in all seriousness, how do you go from being a competitive swing dancer to a data center junkie?

Well, you know, a little, little sidebar to start the conversation. When I was in university in upstate New York, I asked this girl Karen out on a date and she's like, oh, I want to go dancing. And so we went to this swing dance class and I got hooked. I never saw her again after that first date, but I always give her credit. And I just fell in love with dancing because it's a— there's a community aspect to dancing that I just find so compelling that it doesn't matter if you're a parking lot attendant or an investment banker, you know, everyone who goes out to enjoy partner dancing, generally you only go if you're in a great mood to begin with. So you're just surrounded by this positive energy all the time. And it's this real society equalizer in that you all share the same passion. So yeah, you know, I did that more as a hobby for a long time. And I got to a level where I kind of had to make a choice. It's like go down the path I've been studying in school and, you know, I've got a career in or kind of go professional as a dancer. And I'm very happy with my choice that I kept it as a hobby, as like a release from day-to-day stress instead of making it my job. And instead I focused on, you know, technology and I've been in the data center space since 1996.

So someone from my own era. Now I wore this shirt in part in honor of your swing dancing. Kind of background because this is something that I might have worn with a cute little skirt to go swinging myself. So, so you work in this space that we can colloquially call big data and big data scares many people. So can you explain what makes big data so important for all of us do-gooders?

Yeah, there, there's a lot to unpack there. Uh, I'd say the first thing is if you look at the volume of data, that's the important thing to start with. So if you take, you know, probably since the beginning of mankind until today, 90% of all data ever generated has been in the past 2 years. So if you think about the scale it took to get to where we are now, it's very, very impressive. And we're probably going to hit 5 times the amount of data we have today within the next 5 to 6 years. So this growth in data it's both consumption and creation as well. You know, people making YouTube videos or TikTok videos. And so this massive volume of data has a lot of impact on both society and how we operate, but also the environment, what that means from water usage, power consumption, noise pollution, you know, other environmental issues. And, you know, that's really kind of near and dear to my heart. I've got two kids, you know, two young kids, and I wanna make sure that there's a planet here for them to live on.

So something people might have heard about recently in the last few years is the carbon footprint of their, their digital world, so to speak. So even sending an email and storing it in the cloud versus sending as an attachment can have an impact. And so I know people in the space of sustainability that are working to have everything in the cloud and then just send people a link to where that piece exists as opposed to ever sending an attachment. And this can also save the inboxes of people as well and irritate them less because you're not then dumping a bunch of files onto their computer that they may or may not need. So what other steps might you encourage people to employ if they're also looking at, you know, reducing their carbon footprint per se, if it, as it relates to data?

Yeah, I mean, people, it's, it's hard to connect the dots because when we hear data, it's this whole virtual thing and you hear about the cloud and they make it seem very kind of light light and easy and, you know, but the facts of it are very different. You know, if you think of data being stored on your computer, which has a hard drive, there's a lot of minerals being dug out of the ground in different parts of the world. And then you think of the transport cost of that, not in dollars, but in terms of environmental impact. And those things all add up. And, you know, you might have the same file, like in your example, 20, 30, 40 times unnecessarily. So number one, I agree with that. Send links wherever possible because you're reducing the things we dig out of the ground. But it's also how we interact with technology that plays a big part in it. And for my sector, which is the data center side, you know, if you think of the usage of all these technologies, social media, etc., the more we use them, the more power is required. So in 2016, globally, data centers used about 250 or 270 terawatt-hours of electricity. And by 2030, so a 15-year period, is gonna be about 4x the amount of power. It'll be about 1,000 terawatt-hours. And so if you think of everything required to deliver that power, you know, you have to have power generation facilities that are natural gas or coal, maybe they're solar or wind or hydroelectric. There's a lot of options. But then you think of the transmission and the distribution of that power and how that impacts the communities where that power goes. You know, a lot of times it's a net negative for these communities. And so I think it's, you know, it's hard to convince people to change behavior to be good for the planet. And that's where, you know, my focus is trying to make sure that everything behind the curtain has improvement so we can reduce the burden of change on consumers and absorb more of that change as the business.

So where are your data centers located around the world?

So our facilities are designed what's called edge computing. And so edge computing, think of a big data center only much smaller. So same capabilities, but, uh, we're doing them in form factors that we can make them inside of a factory. And by making them inside of a factory, we can produce them much more consistently. We can have transparency on the embodied carbon that goes into making a facility, and therefore we can deploy them anywhere we want to deploy them. So Redivider facilities are what's called edge computing, and these are everything you think of a big data center, only in a smaller form factor. And we've kind of chosen that model so that we could deploy them in thousands and thousands of locations. But in doing so, you know, the idea is to start with this blank sheet of paper and how do we do this in a far more sustainable way? How do we do this with higher social impact? And so that's led us to the model of making them prefabricated. So we're creating them in a factory. We have higher consistency on what we do, which lowers the cost, which is great. But also importantly, we can measure the embodied carbon that goes into every single facility. And because we're also putting them in locations, uh, wherever there's a need, right? So, uh, generally near where population centers are and, uh, traditional data centers go where population centers aren't, right? Because they're very big, loud, airport-sounding facilities. So by putting them closer to population centers, we can increases sustainability. We can also do what's called higher chained GDP, where we can hire locally, we can train people locally, we can have a more heavy hand on the impact that data centers would have on any individual community.

So beyond what you're already doing to reduce, let's say, that carbon footprint, um, making them accessible to a particular community so that the data is traveling less, so to speak, even if it is in the cloud. Right?

That's right. That's right.

And what else are you really doing to benefit people and planet, not just profits?

That comes down to a range of different things. So we've decided to really focus our alignment on the United Nations Sustainability Development Goals. And the reason for that is companies big and small around the world have all made these pledges for either net zero or carbon reduction targets and things like that. And the Sustainability Development Goals are a great framework that are universally understood by corporations, but at the same time, they're very difficult to build a large enterprise plan and work towards those goals. And from a computing perspective, this is a real easy thing for us to showcase. You know, today you're running compute power like this. By switching to an edge form factor that Redivider offers, we can instantly help you on the, you know, these SDGs, right? SDG number 2 around zero hunger, because we can take the waste product of our facility, which is hot water and hot air, and use that to help deliver greenhouses for these communities to change the relationship children have with food, for example. Quality education is another SDG Number 4, and, you know, in these small communities around the United States that we're focused, which are called Opportunity Zones, we can really focus on improving the skill set of that population to align with the jobs of tomorrow instead of the jobs of yesterday. And just by case in point, our objective is to train no less than 25 times the number of staff we hire in any one location. So the path that we've taken is to take this growth of humanity and this growth of data and really shrink the gap between them, which has historically been getting bigger and bigger and bigger. And, you know, that's a powerful basic idea, and it's one that we think the entire industry of data centers will adopt.

How do you see AI coming into this world in particular? I know it's of the moment to be thinking about the implications artificial intelligence is bringing into our world, specifically as it relates to technologies and even in the green solutions that we're working to create. So how do you see that changing or supporting what we're doing today in the next few years?

Yeah, I mean, AI— I tend to use this catchphrase that 2022 was the last year that humanity existed in isolation. You know, AI is forever going to be part of the fabric of humanity here forward, and the implications on that are really massive in terms of productivity gains or even how we solve problems, not just the problem solving itself. But again, the computation required for AI is incredible. The environmental impact of AI is incredible. And so that water consumption alone, you know, if you're using ChatGPT, you know, if you have any sort of short conversation with one of these AI platforms, you're more or less dumping out a half a liter of water every time you do. And so trying to come up with solutions that reduce the dependency on water consumption for data centers is important.

And I think that people will have a hard time understanding why data centers consume water if they don't understand the heat exchange issues. So can you briefly describe why this is important? And then I might draw another example to help people visualize this as well.

Sure. Yeah. I mean, computers generate a lot of heat. And as computers get more and more dense, they're generating a higher density of heat as well. So typically that cooling is done with air or these things called a chiller, which basically takes hot air and uses water to evaporate cold, have a cold air system that then goes back into the front of the computer and hot air comes out the back. So that process is extremely inefficient. You know, you're, you're only able to recycle about 30% of that water. So the majority of that evaporative process just goes back into the atmosphere. So we're taking the fresh water that could be used, like the city of Phoenix is a great example, is putting a moratorium on residential housing in the future because they don't have enough water. Well, guess what? There's tons of chip fabrication plants and data centers and everything that's down in Arizona. And we're now competing for resources. Are we going to drink it or use it for compute power? And that's a place we really don't need to be. You know, there are other solutions that we can use and employ behind the curtain of the data center industry to make improvements there.

Yeah. You know, my husband and I, back in 2010, we flew to Australia and New Zealand for a whirlwind vacation. And he worked at the time with a company that utilized Weta facilities, which did all of the computer graphics animation processing for the Lord of the Rings trilogy, right? And we had an invite to go visit their data centers and see it in action. It was one of those that had been created to be entirely naturally water-cooled with kind of like a looping system. And then it was built on this floating space so that if there was ever a flooding issue, like something cracked, it would go underneath the computing system as opposed to damaging it. And, you know, so I heard about all this. I'm like, wow, I never really knew that we would use water and cooling. And then I make that into a microcosm in my own home. My husband's a computer engineer. He operates a server out of our home. We've placed that server in a closet that has extra vents in it that is bolted literally to the cement foundation of our lowest level because it's coolest here, right? And you know, of course it has fans running and things like that, but we're just talking something that could run maybe a small company, like throw 5 or 10 computers at it and you're good. But that consumes nowhere near the energy of what you're talking about, creates nowhere near the heat of what you're talking about when it's one of these data centers. And so I think the listeners who may not be as familiar just really need to understand, imagine the supercomputers of, you know, the early days when a single, you know, 386 computer would probably replace that first computer. These are large centers, right? There are a lot of computing devices happening at once, storing data, and you need to find a way to cool that and efficiently.

So yeah, an average data center would use millions of gallons of water per day just to have the unit economics be understood. You know, we're talking a huge volume of water. And that's, you know, that's one of the environmental concerns for sure is water. And the same thing with the electricity, same problem, kind of different way of looking at it. When you have a power station, they're generally not near population centers and they use the wires that you see in your neighborhood as they're called transmission and distribution lines. And the amount of electricity you use, sorry, you lose is dependent on how far away you are from those generation facilities. So you're now taking this other environmental issue or fossil fuel to, to generate power. And now you're throwing at least half of it away before it gets to a baseload facility like a data center. And so when you think, okay, well, the water in the data center is bad, but now you have all the water at the power generation, which has the same cooling concepts. It's the same system. So you're actually losing water at two major steps to generate the data so that you can use TikTok on your phone, right? And so water is definitely a massive issue, uh, for data centers. And, you know, as it becomes a more scarce resource for humanity, we just need to think differently about how to do cooling for massive power consumption, things like data centers.

So when you're talking about building these data centers where they're needed, it sounds to me like you're talking about building smaller data centers, like not necessarily the massive buildings. Is, is that accurate, or are you— how are you so different?

So yeah, that, that's right. I mean, a, uh, a really large facility probably is a unit of measure today, uh, is about maybe 300 megawatts is a, is a large data center. You know, 20, 30 years ago, it would have been 5 megawatts that was a large data center. Now it's 300. We think that the opportunity to address sustainability and social impact is edge computing, which are smaller facilities that can go closer to where the source of the data is, right? So if you think about it, data is generated from somewhere and it's consumed from somewhere, and the shorter you can make that distance, the more efficient you can be across all of these concepts. And so by making facilities smaller, we can put them closer to the 5G tower, for example, So you get faster speeds or they're closer to a smart city, which doesn't want their data to leave their local area. And so all of our facilities are smaller in size. They, they're modular. So you can make one, you know, hold 10 racks worth of servers and you can make another one that holds 100 racks worth of servers, but we're not designing any of these to be the size of football fields.

Good to know. You know, I grew up essentially in Silicon Valley and I've even been a part of recent protests in the heart of it, Palo Alto, right? Seeking to hold the city more accountable for their, the carbon footprints that they'll even allow with building codes and things like that. Right. And in many cases, those living in Palo Alto are still encountering quite a bit of resistance along the way in the interest of business. Because even, you know, this uber-liberal space is very, very business-friendly in many ways. So while building greener solutions might be desirable, the powers that be often bend towards those interests, the interests of business, and not necessarily the will of the people. People might want convenience but not understand the cost at which that convenience comes. So when do you think the market will care enough about sustainable solutions and and catch up with the will of the people to live greener and more sustainably?

I think that moment is right now, to be honest. That's the key unlock that we discovered, which is why we created Redivider. We started in 2021, is when everyone started working from home, right? So if you think of how it was pre-pandemic, you had a very large office building with a huge internet connection to a very large data center, and that made that connection very kind of simple in concept. Well, as soon as everyone started working from home, this decentralized nature of work that is here to stay, it changes the fabric of how we use data and move data around. And so the environmental footprint of that also changes significantly, some for better, some for worse. But that was also the moment where it's like, aha, this is the moment for edge computing to actually deliver on its promise. And the adoption of higher sustainability practices, again, I think is companies generally care less about that than they do profits. And so the way that we're tackling this problem is to say, well, you already have whatever X computing resources that you need, and you're going to, because of AI and IoT and smart cities and, you know, all this adoption of technology. You have these massive growth plans as a company. It doesn't matter which company it is. They all have massive growth plans. So you can either spend the same dollars doing it the old way, or you could spend it doing it our way, which automatically gives you higher sustainability, social impact, and alignment to these UN sustainability development goals, which are at the board level, right? The board is caring about moving that down the road, and the IT department is more concerned about their requirements. And we think now is the time that effectively both of those groups inside of an organization can win together.

I mean, it sounds like the demand for data in this decentralized way is really creating a strain. I mean, I'm, I'm going to point to something that has happened here often over the course of the last couple of years. You know, I'm in Scotts Valley, which is in Santa Cruz County, which is basically a bedroom community of commuters who work in Silicon Valley. We've had a massive movement towards people, you know, especially that work in tech. Everyone wants to drive a Tesla, right? And so PG&E has actually made statements which are now buried. You can't even find them if you try. But they've made statements like, well, you can have air conditioning or electric cars, but not both. Okay. So we literally have had rolling blackouts that are not announced. It's just like suddenly, hey, the whole system's, the grid's down and you might get power back tomorrow or in 5 hours, which is incredibly disruptive to people who work from home and are now relying on those connections to be able to do their jobs, right? You— I have a generator. I've run it more often than I'd like. Yes, I have solar, but I'm not off the grid because I was advised when I installed it to feed it back to PG&E up the line, right? And when PG&E is down, that doesn't happen either. So now I have to, as this dark green individual who does not like this fact at all, crank my generator, which runs on natural gas, to, you know, be able to work and function. I had to do this for as long as 5 days when we had a massive power outage due to the wildfires that were coming through when we had high winds. So I imagine that these sorts of things, these infrastructure challenges, also impact data centers. I wonder what your thoughts are about what challenges could erupt that we haven't seen yet. You know, if you were to be predictive, perhaps, like, how can we change these infrastructure challenges into opportunities that we can solve in the near future?

Yeah, I think, you know, your story is very analogous at a personal level of what's happening at an industrial scale. And we think that the solution to that is to disregard the grid. Now, obviously, that's a big statement to say, just stop using— we need to come up with new solutions to the grid.

Microgrid or something else, right?

Yeah. Yeah. It's going to be very transition, like a transition period to do that, or think of the grid as a backup, but we call it prime onsite power, right? So you, you use a microgrid island for things like a data center that are considered a baseload power customer and anything that's baseload that's running 24/7, its primary power, it should be responsible for its own primary power. And that's what we're, you know, focusing on next. Having built out the supply chain for prefabricating data centers, we're now embarking on doing the same exact thing for Prime Onsite Power. And that does 2 or 3 things, right? It does not put strain on the grid today because that should go to houses. It also ensures very high-quality uptime. And it also allows us to do the same sustainability initiatives on the data center but do them for power generation.

Awesome. Now, another question I have for you, and you may not have an answer for this, it was just a curious thought I had as I was preparing for this interview. Um, as I was reading Paul Hawken's Regeneration: Ending the Climate Crisis in One Generation, I had actually stumbled on this whole concept of heat pumps and how it's a more efficient means to both heat your home and cool your home. And I wondered if there, that technology had been applied to data centers, if it's similar, because it, you know, we're wanting to recycle, reuse, you know, not lose the water. Is there a solution that potentially can draw from that technology?

Absolutely. I mean, it's funny you mentioned the Tesla. I drive a Polestar because it's the most sustainably produced car ever made, and it has a heat pump that has very high efficiency, and it's got carbon tracking throughout the entire, every aspect of the vehicle. You should look into the Polestar. It's a totally different mindset about how to build a car. But yeah, so the, the idea of heat pumps exists in the data center space, and there is a huge desire for this stuff. The issue becomes that desire falls second to meeting the demand of 5x data growth volume needs, delivering on the speed promises of 5G, the fact that we need 4x more power in the next 15 years just for data centers alone. And that doesn't even include AI and IoT. That's not factored into those numbers, right? So People are, you know, my prediction, unfortunately, and, you know, I'm obviously on record saying this, but I think within the next 10 years, America is going to start turning coal plants back on to meet the demand of data growth. And that is horrible. But I think that's the reality of what society is choosing. We're choosing TikTok over Mother Earth at a very primitive idea. So the idea of heat pumps and all these desires are there. But they haven't been put into action because it's faster and cheaper to go another way. And that's where we think edge computing and starting with that blank sheet of paper to do it sustainably from the start, costs are identical, but this one aligns with your corporate desires of SDG alignment. The traditional way doesn't. We can do prime onsite power or we can use the grid. There's a lot of great things with natural gas. That are not yet kind of in market. But, you know, hydrogen, the hydrogen economy is here as well. And for things that are baseload customers like concrete or glass or data centers, you know, the idea of the hydrogen economy will be fundamentally different for how we, we build the future.

Well, I think there's obviously a lot of opportunity here for, for growth in the right direction. I've also wondered when we're going to turn off those coal power plants. I was recently pitched for this podcast to cover the topic of nuclear power, and I'm still kind of waffling on whether or not I want to open that can of worms because it feels like a lot to get smart on. I've heard people in the eco-friendly space say that nuclear power can be greener power, but I think solutions need to come from other angles as well, and The problem I see is as we, you know, rely more and more on data storage and also on energy storage, rare earth minerals are going to become increasingly hard. I read recently that Norway found one of the largest stores of magnesium underground, which can be used to create energy storage capacity. And so it's quite possible that that will be worth trillions upon trillions of dollars as technologies continue to go forward. It's like they've just discovered oil again. In Norway. And so, you know, we're kind of, in my mind, in one way we're robbing one resource to pay for another. And I'm not sure where that story will land us long-term, but it feels like we need regenerative solutions that aren't going to degrade our environments and where we don't have to dig into the seafloor to get to them because we don't know what the impact on ecosystems will be in that case. I mean, there's even some that are saying they're, you know, well, we think we can make more noise underwater than we presently are, but then advocates of marine I'm paraphrasing, but that damages the sonar frequencies that these whales rely on to navigate and sends them beaching onshore and things like that. So there's— that's exactly it. There are— there's so many ways in which a single choice has repercussions that we don't necessarily always go through the work of ferreting out before we make a decision and plot a course. So, I mean, there's no easy solution in the near term on this stuff, but I really hope that we work harder as a people to think about what the long term is going to be in each of these cases. If we choose a course, what does it mean? And, you know, if we choose to, to live the TikTok world, what does it mean? What does it mean to our shared connection? What does it mean for data centers? What does it mean for storage? What does it mean? I don't know. I like the platform too.

These are, these are really good points to talk about and just to chime in on that. So for base load power consumption. Yeah. So these are really big issues to talk about and they're, they're very important and there's not a universal solution here, but I will say in the domain that I operate for the base load kind of power customer of a data center, the solutioning gets pretty simple to focus on sustainability. It's not going to be solar. Right? Because solar, just forget the time of day thing and all the things that go along with that for batteries and whatever. The land clearing required to have a solar installation for a data center is dozens of acres of land per 1 megawatt, like recurring power. So that's just not viable.

You can't do it with a Tesla tile roof.

Yeah, it just doesn't cut it, right? Solar is great for a house, you know.

I know.

I think it's a great solution for residents, but I mean, yeah, it just doesn't— solar doesn't— it's not the, it's not the silver bullet. I mean, you could put solar on every single house in America and it's still not going to solve the core issue of power. We'll still rely on coal. I mean, if you look at renewables, you take the entire power consumption of the world as a pie chart over 20 years, and I forgot how many, like $500 trillion of investment. it's only gone to 2 or 3% is renewable, right? We still rely on fossil fuels for a lot of things. And the baseload power portion of that just for data centers is also about 3% of total power consumption. So we need solutions that aren't solar. Wind is not a viable option because it's very proximity-based. The lifespan of these turbines is like half what they thought it was. Like, it's just not a great— and the CapEx on wind turbines too high. So that's just not a viable option. So, you know, we really think that it either needs to be a massively refined process for natural gas to make the fossil fuels that we're going to consume anyway, make them have less of an impact, or we need to focus on hydrogen. You know, we think those are the two real big opportunities to do this in a new way that actually moves the needle. Otherwise, we're just, like you said, we're actually transferring the problem. Yeah, you're transferring the problem, but even more so, it's now a geopolitical risk because the majority of these resources we're digging out of the ground are not in the United States. And so you take this macro trend of deglobalization and depopulation, and you overlay that with the fact that countries who have these resources are going to start going, hang on, maybe if the world needs these so bad, maybe I'll keep them for myself so everyone else can have the depopulation problem. You know, that is going to be a moment of reckoning for every country in the world. And hydrogen is the one thing that I think universally could address this problem. But it's just, you know, it's a new economy to build.

It's a new technology, newer, right? But there are even hydrogen-powered cars. I think BMW has been doing quite a bit with them. And you know what, the emissions are water. So that's, that sounds pretty darn good to a lot of people, right?

Yeah. The hydrogen problem is it's it costs more to make hydrogen than it does just to make the electricity that would've been the consumer. So there's still some evolution that needs to occur to have that be viable when you're talking about 4x-ing the power consumption of data centers alone.

Well, the other one, an interesting technology that I learned about years ago when I was sitting next to somebody on an upgraded business class flight, right? And this was a gentleman who was spearheading technology to harness the power of ocean waves. By basically having these balls on the ocean that would go up and down with the waves and it would create energy within them. And I was like, well, how does that get networked and then get to the shore and not impact marine life? He's like, well, it actually becomes a living reef and giving me all what sounded like really great answers. But I've never actually seen the technology take off or go beyond that kind of pilot phase. So I think if we're able to reinvent, you know, something like wind energy, I've seen those that are low to the ground that kind of take more of the trade winds in, a tulip shape and things along these lines, which can be placed even in city centers and things like that might also help. But again, it seems like it's a lot of infrastructure to build and try to connect. Others that are using the natural heating and cooling principles to go ahead and store and release energy. So I think we'll, we'll see where these end up.

But, um, yeah, geothermal is one I'm very, uh, excited about. But again, I think that's, uh, you know, measured in decades, not days. But, you know, if you think of the Earth as a battery and you think of outer space as like, uh, you know, a cooling system, there's some very novel technologies being developed at university right now where they're using a very narrow spectrum that can essentially access the cold of space to be a heat sink, right? So if you can solve some of the cooling problems, then you can reduce the power consumption as well. So it's going to be a concert of these technologies that come to fruition that allow us to make meaningful change. The issue is, are we going to care? You know, is humanity going to care enough to make a change? And if we can't, like, again, the perspective of What we're doing with Redivider is just to say we'll do all that alignment, we'll build, bake it into our solutions. So you have these growth plans, they're going to cost you X dollars whether you go left or right, but if you go left, you're doing all this other great stuff. And we think that that is a cheat code for big companies and small companies alike, and at least for the lane that we're in, which is data, right? All the other things are, are still problems to solve for.

Big discussions. Now, geothermal plants, I think the only country that has that thoroughly dialed is Iceland. And I happen to work with technology that's co-located with a geo— the one of the world's largest geothermal plants on power in Reykjavik, Iceland, right? And so we're taking their waste stream CO2 to actually grow algae that we can then build into nutritional solutions and supplements. And that's what I do for my, my day job, right? But you don't have that situation. Everywhere around the globe. Like what, what are we gonna do suddenly go to Yellowstone and, and put a geothermal plant there in the middle of, you know, what is this natural preserve? You know, that would not be ideal. So we really have to, that resource is limited and in some cases you simply can't build a power plant where you have access to the, that geothermal energy. So I'll stay tuned on what that technology is bringing out, but I, I would love for you to be able to share with our audience. Really what can we do to get businesses to change data center practices for good beyond, you know, I mean, I don't think I can affect Google, but Google already claims to have green data centers as it stands. So I think they're doing that through carbon credits mostly, but I'm curious to hear what your thoughts are there.

Well, specifically about Google, I would say they are by far leading the market. They care, you know, and they've been at it a long time and they've done a lot of experiments that have failed, which is great in terms of progress making things, you know, you have fewer things to fail at. So I would say that they have done a great job. As an industry, what can consumers do to help shift it? I'm not sure it's necessarily a consumer solution that's needed, right? I think it's about kind of the B2B solutions that are required. And a really easy example is 5G, you know, the all the consumers are spending to get faster speeds, but the new speeds that are being delivered is largely from these 5G antenna that have much— you need a higher density of antenna to deliver on that speed. And so the idea of the telecom industry partnering more closely with the data center industry is what's going to make that meaningful difference. And that's why we think the The interim step right now are what's called the cloud edge. So take Google and Amazon, SAP, Salesforce, Microsoft, et cetera. They all have these clouds. And by putting a smaller version of their cloud closer to customers that can also deliver on these 5G speeds and the data doesn't move around as much, you can effectively lower that carbon footprint. You change the water consumption of data on a unit basis. And, you know, so ultimately I think that the higher adoption of 5G by customers would impact the relationship between telco and data centers that could deliver that kind of result. But in terms of other things, you know, I think you nailed the number one thing that I was going to suggest for consumers, which is share links, not files, because the, just the rare earth minerals required to make a hard drive at scale for all of the collective cloud is a huge number. And a link means that file exists in one place. So I'd say that's the number one on the consumer basis, but to really solve this problem, the core relationship that needs to be addressed is telco to data center.

Telco to data center.

What is telco? Telecomunications.

Okay. I'm sorry. I heard that as toco. I was picturing my big toe.

Telco. So the Verizon's—

Telecom was, I think, how I heard about that. Telecommunications. Okay. That's, that makes sense. More sense than what I was thinking. Great. So at this point in our conversation, Tom, I'd just like to ask you, you know, what parting thoughts you would like to leave our audience with, or if there was a question I haven't asked that perhaps you wish I had, you could ask and answer it.

Uh, I think the only other major thing to, at least relevant to what we're doing, is the idea of the construction workforce and its impact on communities. So traditionally, when data centers get built, they use large national contractors who come into a community, they work for 2 years, and then they leave. And the economic impact that that has is long-term low because they're not hiring locally, they're not doing all this stuff locally, it's all done nationally. And now we're seeing this massive atrophy in the construction workforce, right? Since COVID happened, we're seeing you're not getting new journeymen come in for electrical, you're not getting new tradesmen coming in for building, and that's going to push out all of these construction projects for data centers from 2 years to 3 years to 4 years to 5 years. While at the same time the demand for data continues. And the impact of that, what that's going to mean, and I alluded to this before, we're going to see dirty technologies turn on again that haven't been used. We're going to see an increase in the coal usage for humanity to use Facebook, right? And that's directly correlated to this atrophy in trade workers that are building things. So there's really two solutions to that. One is prefabricate them, which is the path that we're on, or two, America needs to get together and collectively decide that having a trade job is awesome and we need to figure out ways to compensate people for that and reinvigorate that entire industry. Or I do think we will see a net negative environmental impact.

Wow. Well, I will direct people back to my interview with Paul Hawken from almost two years ago now, where We talked about the retraining of people who are in specific trades to be more on the green technology side because it's what we'll need. And I think you're speaking to exactly the same thing. I'm impressed with what you're doing with Redivider. And I think that we should stay connected because as your business unfolds and develops, I would like to connect with you again. I really think you're doing important work, even if to the layperson it might sound a little, you know, background and perhaps boring. These are resources we all use every day. You can't get away from your cell phone at this point in time.

Yeah. And again, it's only getting bigger and bigger and bigger. And, you know, I've got all these statistics I can share with you and, you know, where I'm pulling these numbers from. But the big asterisk is all of that, those stats are pre-AI, they're pre-IoT. So when you factor in the largest trend we've ever seen in computing since the 1960s, Everything I've shared with you, I think is massively sandbagged. You know, I don't think it's going to be 5x data growth.

This is, this is the next frontier. That's where we really are. We're on.

Yeah.

It's the beginning of things. So I will also point people to listen to the episode I recorded with Mo Gawdat, who was the chief business officer for Google X for many years. He wrote a book called Scary Smart, all about AI. And I re-released that episode a few weeks ago just to make it easier to find along with an intro commentary on AI and where we're headed, because I think it's doubly important now. We teach AI that we can be quite nice, and it's not all, you know, bad when you go to the likes of Twitter and see the trolls and everything else. I mean, we can be quite kind and impact-driven individuals who care about the greater good, and so we should begin training AI more about what great about humanity.

Yeah, actually, it's funny you say that. I'll just give a last little side story. Uh, when, when all this ChatGPT stuff came out, uh, late 2022, I really had this, like, idea to— I wanted to test it, like, for real, not like a playground, but really test it. And I looked at all the education material for children in different languages, and what I found was, if you think about a language like Ukrainian or Spanish or just Irish, right? If you look at all of these other languages that exist, the education material for kids is less than 1% of what's available in English. So, you know, I built a whole system to design kids' books that is a kind of well-thought-out, like, educational framework, but deliver them in different languages in a way that the language is meant to be used. And, you know, I was blown away at the feedback that I was offering these. I think my little test in December, I published like 90 books or something, which is a huge number to think about it, but AI made that pretty simple. But the feedback on that has been just incredible. So AI has the potential to help humanity do things that have never been done before, you know, in terms of how we, we live our daily lives and learn things. But, uh, you know, it's a double-edged sword. Because it could also do, do bad things too.

Well, and I think again, it's showing the best of ourselves. I mean, Mo Gawdat made the point, he's like, look, if you spend all your time on Instagram paying attention to, you know, hot body challenges and not to when people are communicating with authenticity, if all you're doing is looking at filtered faces and not real faces, then you're teaching AI that that particular trend too. And so even our simple actions that we may think of as kind of mindless, it's learning. And that's the nature of AI. So when we get to the singularity point where it has far surpassed the mental power of humanity, it's not crazy distant. It's going to happen in my lifetime, yours, and probably almost everyone listening today. So We need to be ready for that. And I think we need to think about the far future when it comes to solutions and technology too, which I'm happy to see you're doing. And I just want to thank you for your time today.

Oh, it's great to be here. And, you know, this stuff is near and dear to my heart. And like I said, I want to leave a planet for my kids to be a part of. And at the rate we're going, we're sucking all the life force out of the planet. And it's our responsibility, you know, it's It's Gen X, it's millennials, it's Gen Z. We all have an obligation and opportunity to do the right thing and do it in a way that is positive for people, planet, and then ultimately generate higher profits.

Yay. Here, here. Here, here. I'll raise my cup of coffee to that. Chin chin. Sustainably harvested. My friend, Sprout and Blossom Coffee, it's low mold and really, really good. It's delicious. But whatever you're drinking, whether it be water, glass of wine, Cheers to you, Tom. Really great to have you here today.

Thank you. Appreciate it. It's great to be here.

To learn more about Tom Frazier and his work with redivider, visit redivider.co, spelled exactly like it sounds. I'll be sure to include all direct links to the resources that we discussed in our show notes. Just visit caremorebebetter.com. That's exactly where you'll want to go to get a closer look at today's discussion. You'll even find complete transcripts, related content, including my earlier interview with Mo Gawdat and also Paul Hocken, who are both New York Times bestselling authors. If you liked today's episode, please subscribe and set that bell to notify you when new episodes drop. And please also leave us a review and tell us what you thought of today's episode. This simple act will help more people discover this content. Thank you, listeners and watchers, now and always, for being a part of this pod and this community. Because together, we really can do so much more. We can care more. We can be better. We can even create that greener, better, easier future with responsible data, edge computing. Because for all the necessary evil that they may be, they are also an incredible tool and asset. To connect us and build a little bit of ease into our daily life while also entertaining us. Thank you. Thanks for listening to Care More, Be Better, a podcast for social good. To make sure you never miss an episode, subscribe, rate, and review wherever you listen to podcasts and share with your friends to help us reach more people and spread more social good.