Become a Creator today!Start creating today - Share your story with the world!
Start for free
00:00:00
00:00:01
Can the tech industry deliver on AI and clean energy?
77 Plays1 month ago
This week's discussion is all about the tech industry. Karthik, Anthony, and Amy first check in with podcast standby Elon Musk, whose recent demonstration of "AI"-powered robots and robotaxis was more flop than futurism. Next, they turn to tech's investment in nuclear power. Companies like Amazon are committing millions to nuclear, but it's going to take more than piles of money to make it work.
Recommended
Transcript
Introduction to Innovation Matters
00:00:12
Speaker
Hello and welcome to Innovation Matters. It is the podcast about sustainable innovation brought to you by Lux Research. I'm Anthony Schiavo. I'm Senior Director here at Lux Research, and I'm joined by my two colleagues, Amy Hurus and Karthik Subramanian. Amy, how are you doing today? How are things in beautiful, warm, sunny Africa?
00:00:32
Speaker
i'm Great, not so warm. It's still funny though, but we have in Ottawa, the first below zero temperatures and that's all Celsius, of course, but yeah, the winter is coming. Yeah, we have not we have not gotten down there. It dropped. There was a noticeable step change downward to about 40 degrees Fahrenheit, 45 degrees Fahrenheit um here in Boston, but we haven't cracked the freezing threshold just yet. Not yet.
00:01:01
Speaker
Karthik, how are you? I'm good as well. um think we I don't know if you would call this a heat wave-ish. I would call it a warm wave because the temperature suddenly went up to like 21 degrees from 13 degrees out of nowhere, oh um which is very unusual for ah you know almost the end of
Health Reminder for Male Listeners Over 40
00:01:18
Speaker
October. so Mike cannot join us today and he has a special message to all of our listeners, particularly the male listeners in our audience, which is that colorectal cancer is no joke. If you're over the age of 40, you need to get screened, even though it's uncomfortable. And Mike's commitment to this type of of public health is so substantial.
00:01:40
Speaker
that he himself is currently, as we record this, undergoing that screening. So remember, fellas, if you're out there, we know we've seen the analytics on who listens to this podcast. We know you're a little bit older. um Don't kid yourself.
00:01:58
Speaker
you're not special, you do need to get you know get the screening done if you're of a certain age. So that's just a little Innovation Matters public service announcement. And speaking of men who are of a certain age, it's Elon Musk once again.
Critique of Tesla's Robotaxis and Optimus Robots
00:02:12
Speaker
We we are we are returned to to talk about our our favorite guy on the podcast. um We have a couple of things we want to go over relating to AI and tech companies and a couple other things. But we wanted to start with with Elon and his recent event. He did a Tesla event where they showed off a robotaxi as well as their quote, you know, Optimus robots.
00:02:40
Speaker
um These are their humanoid robots, if you haven't somehow blessedly managed to miss this, good for you. you the They basically did this event where they showed off the robotaxi and then the the robots walked around and served drinks at the at the event.
00:03:01
Speaker
Elon made it seem like they were you know autonomous. He said they're the biggest deal, the biggest product of all time, I think is what he said. Of course, it turned out pretty much immediately that the robots were being remotely operated, which is which was pretty obvious, I think, to anyone who was familiar with the state of the technology.
00:03:21
Speaker
um But you know I think the overall reaction to this, certainly from the markets and from general outside observers, was was pretty negative. ah The robotaxes themselves, Elon said they were going to be ready by 2029 for mass production. um He's basically missed every single target he's ever set, so no one particularly believed that. um They also looked a lot like the Cybertruck, which maybe was part of the negative reaction because they're very ugly.
00:03:48
Speaker
ah And I think there's still a lot of skepticism about the robot taxis as a business. you know Waymo is, in some senses, successful, and this is something I want to touch on. I mean, they're doing hundreds of thousands, I think they've cracked 100,000 fully autonomous rides, but their approach is very different. They have these cars covered in sensors, um they're pretty clunky looking, and there's still a lot of challenges with the business. you know again There's a lot of human intervention. right Those cars tend to get stuck. They are being overseen by a fleet of operators. you know that There's still a lot of intervention. And so I think people are really skeptical if Elon can create a ah pretty big step change here. I'll share my thoughts in a moment. but
00:04:33
Speaker
First, Karthik, what were your thoughts on this whole thing? i Looking at it and seeing some of the reviews because I couldn't see the event live. I was just thinking about this and I was like, they don't have or or or seemingly they don't have an override switch of sorts. It's apparently in the display.
00:04:52
Speaker
of the device, um um even before we get into like you know timelines and regulations and whether they can get it online by 2027, which is what Musk claims they will be able to do, and at $30,000, which also seems yeah a long shot. ah But just from um an operational standpoint, i mean I would be scared if I am caged up in you know in a box, if you will, if you could call it that, because it does look ugly to an extent, with no way to get out, because there are no latches on the doors, there are no buttons, emergency checks, there is no way you can control
00:05:27
Speaker
um The autonomous vehicle um if there is an error and they have been notoriously known for you know Missing commands and going into areas that are restricted and I believe there was one Robo taxi which went into ah a Wrong part of Chinatown during Chinese New Year and people started breaking the whole thing up I wouldn't want to be stuck inside that thing if something like that happens with no way to get up You know, I would be super scared So yeah, I mean, that's already an issue in the Model 3. So at the Model 3 back doors, they have an electronic button that controls their door. So if you are in a crash and the electronics fail, you cannot open the door. You have to there is a physical switch, but you have to like pull part of the panel off the door in order to access it. um It's like insanely difficult to to get out in an accident. It's one of the many reasons why I would not be buying a tes the model three anytime soon But to that point, I think if i if any consumer wants to you know ride a robotaxi, I think they would be s skeptical about all of these things. um And I don't think Musk did a really good job explaining some of these finer details.
00:06:36
Speaker
um I think they just focused on the pizzazz of the whole thing. like It was very showmanship kind of you know event rather than actually going to the nitty gritty parts of how they're going to get out of ah you know any dangerous ah issues or or any contingencies. How are they going to convince or even have demos show up that show that if ah you know X happens, we're going to do this to help you get out and it's going to be safe enough.
00:07:00
Speaker
So it doesn't it only reinforces what we already know in terms of this is just a whole bunch of sci-fi. It sounds great, but really there is no way to bring it to the market. Yeah. What did you think about the $30,000 price point? Because I think it's, I mean, obviously it's kind of BS in the sense that it seems challenging for them to produce this new model at $30,000 when the Tesla Model 3 is barely, I think it's still generally more expensive than $30,000, especially with the you know if you add any of the upgrades. like For example, the sensors and the capabilities needed for like full self-driving, that adds a lot of cost to the car. um if you just buy like If you buy the full self-driving package, I think that
00:07:49
Speaker
adds, what, $75,000 to the car? And I think it's just for software. um I think the sensors are already in the car. But, um you know, obviously, they're charging money for it because it's it's a high effort thing. But um to I guess to me, what what stood out about the price point, the $30,000 price point was that the idea was that they're going to sell this to other people. And like,
00:08:14
Speaker
who is going to be buying these fleets of, or are individuals supposed to be buying this? like There's a very interesting sort of vision of autonomy happening here, which is kind of different from, you know if you look at Waymo, like that's owned by Google, I believe, right? And if you look at Cruise, that's owned by GM. So there's big institutions.
00:08:37
Speaker
that are developing the software but also and the hardware, the cars, but also offering the service in an integrated way. And it doesn't really seem like that's Musk's business model here. um Or maybe it is, and he's just saying, oh, look how cheap we can do this. But like like you what do you think is going to happen via the the the business model here?
Tesla's Real Estate Ambitions with Autonomous Products
00:09:00
Speaker
um I think ah taking a step back from just the robotax in and looking at the event broadly, I think Musk and Tesla are really trying to enter the real estate business. um Real estate.
00:09:14
Speaker
Yeah, and and the reason why I say that, and and I'm thinking about this, is mainly because of the level of autonomy that they want to bring with their whole product suite. So they want to bring in robots, which seem like you know they could be helper robots in houses. So that's the level of automation. We already know that Tesla has done their Powerwall with the solar battery, you know that product suite. And now they want to bring in an autonomous vehicle, which is induction charged.
00:09:39
Speaker
So if you want to have wireless induction charging, I think they would also be interested in partnering with someone who can build a garage that allows you to do that. So that is what I think is going to happen. Maybe I'm, and maybe this is a long short thought and maybe I'm completely you know missing ah missing the big picture here, but I think that's what they want to do. I think Tesla at one point want to create this ecosystem of a home.
00:10:03
Speaker
where they're like, it's gonna be completely autonomous for you, you're gonna have autonomous vehicles that are gonna be wireless charged, you're gonna have Tesla solar panels, you're gonna have Tesla battery pack, and you're gonna have the Optimus robots you know that are helping you at home. um and And there, I think a $30,000 price point would make a lot of sense because I think Musk has been asking for a really, I think he has been claiming for a really long time that he's gonna bring out EVs that are as cheap as $25,000, right? I think this was five, 10 years ago when he made this claim.
00:10:30
Speaker
So I think it is going towards that vision. and And I'm just curious to hear your thoughts on whether that is what he's actually targeting, ah because typically with these taxis, you know, auto autonomous vehicles, you first think about fleet operators like your Ubers and your rollers, you know, your taxi operators.
00:10:47
Speaker
I mean, I think it's part of his his vision. I think it's pretty dumb, like, if I'm being honest with you. There's, especially in the context of, like, there's a very strange relationship to labor that Musk has. Like, he's very anti-union. He's very, you know, notably sort of anti-labor.
00:11:11
Speaker
And a big part of this seems to be just like, we're going to take all this, this labor, and we are going to, you know, essentially create these automated systems that allow someone in the global South to remotely do a task that would be carried out by someone here. And like, not to make it too political, but like, Elon is also, you know, now this big Trump supporter, right? He like appeared on stage with Trump.
00:11:37
Speaker
a couple of days before this event. $75 million as well in campaign funding, I believe. Who knows what, because he said he was going to give all this money and then the next day he said, ah, not actually. So typical Elon stuff, pretty flaky. But it's, you know, there's very clearly this political, like he's very clearly becoming a political actor to Elon. That's right.
00:11:57
Speaker
ah Even Donald Trump has said he's gonna appoint you on you know to be Some government position if he wins, right? So like it's hard for me to see this whole focus on automation You know as separate from the the political ambitions Amy I'm curious for your thoughts on any of this oh You know automation and you know, we've talked a little bit in the past about how people might react to these robots but I'm i'm curious for what your read is on all all this and Yeah, I think one of the things that stood out to me um is when he just when he was talking about the robots, I saw just a a short clip of um part of his presentation. he He said they'll be producing products and services. um The cost of products and services will decline dramatically. Anyone will be able to have any products and services they want. um So I think he's
00:12:52
Speaker
These sorts of statements are, are um he's suggesting that these robots have the potential to disrupt our economic system in a way. um So like those sorts of statements are really linked to post-capitalist rhetoric. like it's It's basically this idea that like there'll be less need for human labor at the same time. um So that's going to lead to a shift in the nature of work itself. At the same time, there's going to be abundance of goods and services.
00:13:20
Speaker
um And this this undermines the basic capitalist principle right of scarcity and competition and so like I think his His vision is is like large scale. i don't I don't necessarily see it as realistic. Of course not. But consumers, I think like one of the issues consumers have had, and they repeatedly bring this up when they think about automation or humanoid robots, is is the issue of labor, the loss of jobs right and um human expertise being replaced by um by robots or by machines. and
00:13:55
Speaker
um Elon doesn't seem to be trying to like allay any of those concerns. he's He's like pushing for a future where humans will do less labor. The economic system will will need to adjust as labor becomes more automated. um Well, I mean, I think he kind of is.
Automation, Job Replacement, and Musk's Political Views
00:14:12
Speaker
in so This is where I see the political stuff, because I think a big part of the message, right, around, you know, immigration that we talked about in the election webinar that we did and, you know, with Elon is that, you know, the jobs will be um the, you know, there's immigrant jobs, right. And then there are jobs for for real Americans. And like,
00:14:38
Speaker
we're going to sort of allay this fear about automation with a political push, right, that that limits immigration and and, you know, limits that that job market, right, or the the seekers, I should say, in that job market. And so that's a big part of, you know, it's interesting that you sort of brought that up, because again, I see these things as, as really quite connected.
00:14:57
Speaker
Yeah, but then for me, the question would be if that is the case in terms of just labor. I think what Musk is trying to do is to say, we don't want to have people doing mundane jobs. if If that's what you want the robots to do, right? Like you don't want someone just being, um you know, used to screw in and screw something up. Like literally like a screw, you know, with a screwdriver, like a robot can do it. Like you're not using a brain in any way. And which is why I think robotics and automation is is widely considered as eliminating the need for mundane jobs and helping humans become creative, like use your brains for something else. um But if that is the case, from even an immigration standpoint, even if you take that out, I think, aren't they widely missing the mark? Because at the end of the day, it's going to take out all mundane jobs, irrespective of whether you're an immigrant or a non-immigrant. So you basically have to become skilled.
00:15:46
Speaker
I don't think that the the whole concept is very realistic and I think that the other thing I would say is that you know there's, I think a disconnect here between how complex these tasks really are and what's really possible to automate like if you look at house construction in America. It's been highly resistant to improvements in like labor force efficiency, right? the And indeed it's gone and backwards and in a lot of ways. You know, houses are more expensive to build now than they were, especially on a percentage basis of you know someone's income or whatever. um And so, you know, this idea of like, hey, like these these tasks are simple and thus can be automated.
00:16:29
Speaker
is not, doesn't really hold up in practice. I think we saw this with, you know, in the 2018, 19 era, there was this big focus on cobots, right? It's like, oh, we're gonna, we we can have a robotic arm that costs $25,000. And it's cheap enough and simple enough that we can put it in, you know, in a warehouse, you know, a small business can buy it, right? But the reality is that actually human,
00:16:58
Speaker
even a human who's doing a pretty basic job, like, you know, working at McDonald's, you know, there's a lot of interest in fast food automation. That's actually a really, really complex task. You have to manage a bunch of different components of the hamburger, you have to put them in the right order, you have to do it in response to like changing environment, you know, hey, these new orders came in, like, you know, the simplest, you know, jobs that we have that are sort of looked in terms of Oh, like it's like this sort of derisive thing, you know, you end up you know flipping hamburgers or whatever. That's actually really complicated ah from the perspective of automation and and, you know, in a lot of different ways. And the same thing with screwing in a screw. It's like, well, OK, but maybe you have to recognize that the screw is broken or it's in the wrong place or like, you know, something else is happening. There's a safety issue. There's so much that we take for granted as human beings and just the ability to receive instructions.
00:17:53
Speaker
as a human being is is really important. And that's a big part of why when I see something like you know these robots is and like they're being controlled by humans, I'm like, well, of course they're being controlled by humans. right um And I guess my concern or my wonder is, will they just go to market like this? right Like just having a robotic operator And we're just going to say, oh, yeah, there's someone, you know, we've automated certain simple things like it can walk around, you know, move from place to place by itself or with limited input. But for any task, we're just going to have someone in a, you know, in ah in another country do it or in another location do it. Right. And that's just how we're going to position these. It makes me wonder. Right.
00:18:43
Speaker
um I have a question for Amy as well, and and i was because we are going to talk about the consumer angle on this. Would that level of, ah ah let's say let's call it semi-autonomous, for the lack of a better word, but giving the owner of the robot control? so Let's assume you have ah a fire stick or a remote where you just speak into it and you say, grab me a glass of water.
00:19:07
Speaker
And that goes to the robot and the robot picks it up for you. So its motion is automated, but the processor will have to read your command and then do things for you. It gives a sense of control to the consumer that the robot I am better than the robot. Like there is an inequality there. So I don't have to be scared about the presence of robots around me. Do you think that kind of mindset could help them bring it to market if it is semi-autonomous this way?
00:19:32
Speaker
Yeah, like if it required the the owner to give it commands. ah yeah I do think so because I think when we talked a few weeks ago about um consumers' fears around humanoid robots, one of the things was that what if they um do an action that that harms me? Or what if somebody else can take control of the robot and have it do an action?
00:19:55
Speaker
like unlock my doors or whatever so I think um it it might give an enhanced sense of safety around a robot um being inside your home if there are um mechanisms by which you can ensure nobody else can can take control of that and that you have um the ability to to I guess control its actions more more carefully as opposed to just allowing it to um learn from the the environment and decide on on what it should do based on the context. Yeah, and I think there's an interesting um and point you made about like, okay, what is safe, right? It's like, is someone controlling the robot safe? It's like, okay, well, that's, you know, there's a human in charge, right? And it's often an automated system seen as a higher level of safety, right?
00:20:46
Speaker
to keep a human in the loop but like you know what if this person is you know they're untrained they're unskilled right they're they're working really hard they're like sleep deprived or whatever um you know or they're just vulnerable to some other type of you know, coercion or attack, right? um It's like, what's actually going to, you know, are people going to really see like, hey, like, sure, there's this robot in my house, and there's like a person, but it's like this, this person, I don't know, like seeing into my house, kind of like, it's just a very strange. um I think a lot of what we assume about robots and like, our willingness to accept them into our homes in some, you know, form is like,
00:21:34
Speaker
there's not that level of like surveillance happening or it's not like taking a person into our house. And like we know that like you know the Amazon Alexa is listening to us. yeah And I think most people think it's probably listening to them at all times, right? yeah You still have one in your house because The assumption is that it just goes into a big data bank and maybe there's not like a guy on the other end, like listening to you. You know what I mean? Like specifically one dude. But like with a robot, it's like, hey, like if that thing's on and someone's operating it, like there's a person like more or less telepresencing into your house.
00:22:12
Speaker
Do you want that? To me, it just brings up a lot of challenges. And it's just not that easy to be like, oh, well, you know it's just a robot. So people will accept it. It's like, I don't i don't know if that's true at all, actually. I think that's i think there's a lot of challenges there.
00:22:29
Speaker
Exactly. yeah Privacy concerns, not just safety, but privacy and um whether or not data is being collected or movements are being monitored or things like that. and even just i mean Some of the claims Elon Musk made where were very bold. Like he said, these robots could be a teacher. They can babysit your kids. They can serve you drinks. They can be your friend, whatever you think of it's going to do. and just like
00:22:58
Speaker
considering whether parents would feel comfortable leaving their children under the supervision of a robot, right? Like, so there's there's like technical capabilities, whether or not it's actually capable of doing that. And then there's the the additional layer, like, do people actually want that? Does it feel um like something they would actually benefit from? Or would it just cause increased anxiety?
00:23:22
Speaker
I just wanted to ask you a question. So if Musk goes on to, you know, fulfill his grand vision of a Tesla home, I guess you're not going to be buying that. ah you No, I don't think so. I don't think so. Especially not if Elon Musk has designed it. Yeah. Who knows what it'll look like. I think it'll be aerodynamic, but yeah not good looking. No.
00:23:50
Speaker
Well, speaking of anxiety, we returned once again to our that most anxiety inducing of technologies, which is which is nuclear ah power. And there's been some interesting news. you know This is something that's been gaining a lot of momentum, especially among tech companies. We had a group of banks a couple of weeks ago put out a statement where they were like, hey, we're going to We like it when you invest in nuclear. We'll like underwrite your debt. you know like like Do these deals, please. like It'll be good. And you know just a few days ago, we had Amazon invest um in about $500 million dollars in small nuclear
Tech Giants' Investment in Small Nuclear Reactors
00:24:30
Speaker
reactors. I think particularly X Energy is the company. um And I think Google and Microsoft have also joined all this. I think you know microsoft the the Bill Gates Foundation has already been invested, I think, in
00:24:42
Speaker
in X energy, I think Google is is um you know has dabbled in small modular reactors in the past, you know kind of intermittently. So, Karthik, you're definitely a nuclear correspondent. you You're a Tim Apple correspondent and a nuclear correspondent.
00:24:58
Speaker
um What do you think about this, Karthik? What's the high level here? So ah for those who haven't followed, it's not just Amazon, but also Google actually announced that they're signing the world's first corporate power purchase agreement with Kairos Power for I believe around 500 megawatts of ah small model reactors. Now it's important to know that why this is more important maybe than the Amazon uses because Kairos is the only small model reactor company in the United States.
00:25:29
Speaker
um which has a construction permit for a generation for technology. Even X energy doesn't have that yet. oh um Yeah. And so they're actually only building their non-nuclear demonstrator at this point. um But of course, with the PPA, Google says that in the beginning of the 2030s, we're going to have the first, ah you know, SMRs come online before we go up to 500 megawatts through to 2035. And Amazon and Xenergia are, I think, also trying to scale up to five gigawatts of small model reactors for their fleets by 2039.
00:26:04
Speaker
That's a lot of gigawatts, right? Cause each, each reactor is like in the megawatt range, like a hundred megawatts or so, right? Like. Yeah. So i yeah, so it depends. So X energy is I think building a hundred megawatt. Um, and Kairos is also looking at somewhere around that if I'm not mistaken, but any small model reactor would be between 50 megawatts to around 350 megawatt tops, 300 megawatts, not more than that.
00:26:28
Speaker
So, you know, you're talking about, let's say at 100 megawatts, that's 10 per gigawatt. So you're talking about building 50 of these reactors just for Amazon's for Amazon's goal.
00:26:40
Speaker
Yeah, exactly. And for me, these things sound great. I think the nuclear renaissance for data centers make a lot of sense, especially because one of the things that we have noticed with our data center coverage is that there is a lot of outage anxiety. um Typically, data centers do get their power from utilities through power purchase agreements, ah similar to how Microsoft is you know working with, um I believe, Constellation for Three Mile Island. right So you have a PPA, um and then you're going to purchase power from them.
00:27:10
Speaker
But the current portfolios are not something that suit these data centers sometimes because ah not just with grid congestion issues, but you have a lot of challenges just in procuring power on time. So there is a lot of um you know outages and down times from outages of data centers, which is why you see all of these big tech companies that are hyperscaling the data centers, um looking at geothermal and nuclear as possible options.
00:27:35
Speaker
um Nuclear makes more sense in geothermal because geothermal at the current status is you know very geographically constrained. ah Nuclear not so much. So you can deploy nuclear in places where you can't you know have other other sources going. um So you know very good from that standpoint, but even with this demand, I'm not sure if you know these reactors are gonna come online at that you know magnitude even by the 2040s.
00:28:04
Speaker
yeah That's what I wanted to ask, you know, you have $500 million dollars from Amazon, you have this commitment from Google, is money the issue? It seems like the money at this point is available for these companies, right? Between that, between loans, like, is this really, is this really moving the needle of these, are all these announcements? And like, what would you see as really moving the needle on this?
00:28:29
Speaker
um I think the biggest positive for all of the small model reactor companies is the rise in demand for reactors. Typically you do think of nuclear reactors for large scale power generation, so you're looking at you know grid power, you're not looking at microgrids, you're not looking at industrial you know heat and power, combined heat and power with nuclear. So that demand is great news for all of these developers that want to scale their manufacturing capabilities.
00:28:53
Speaker
but let's and and As you point out, I don't think money is going to be the issue anymore if people are going to not just ease your ability to borrow money as the banks are doing, but also we see that all of these hyperscale you know data center operators are willing to pay a premium.
00:29:11
Speaker
for a nuclear power. So they acknowledge nuclear is expensive and they're willing to pay that premium but to get that reliable power. So money is not the issue. The big challenge for me though is number one, scaling up manufacturing capabilities and number two, fuel. yeah So if you're looking at you know constructing 50 of these reactors, I'm not even sure if X energy can construct one reactor at this point in time. And if you want to build one reactor, yes, please.
00:29:41
Speaker
I was just going to say, how long does it take in terms of manufacturing? Well, right now it takes infinity because it's. But what do you what do you think?
00:29:52
Speaker
I mean, first of all, it's important to know that these reactors are not manufactured anywhere at this point. So they're all bespoke projects and they're constructed. And construction takes anywhere between 10 to 15 years. The world's first SMR in China took 12 years to come online. ah They started construction in 2012, which started producing first power end of December last year. So about 11 and a half years before they actually got it on online.
00:30:15
Speaker
And on top of that, it also you know cost 300% of what they actually estimated to build the reactor to build so it. it became very expensive. And again, because these companies are willing to pay that money, I think that 300% is not the issue. But if you want to get these reactors online,
00:30:33
Speaker
And if they want to bring the reactor online, let's say by 2030, they should have started construction by 2019, if you ask me, which they have not. So given current trends, all of these reactors are going to come online, especially the first of a kind only by 2035.
00:30:48
Speaker
Now, if we assume that these reactors face a lot of delays, which typically happens because of fuel procurement, um especially fuel fuel procurement is a big challenge because fuel comes from Russia. And if you are geopolitically not interested in having ties with Russia, then you need to build your own fuel supply chain.
00:31:07
Speaker
not just you know construct your power plant, then that becomes very difficult. um So you need all of these things to scale up at the same time, and which is why I think first of a kind reactors like the very first unit, maybe the first few units will come online anywhere between 2035 and 2040.
00:31:24
Speaker
But this puts more pressure on the small model or reactor developers than a Google or an Amazon or a Microsoft because these companies are very clear when they sign these agreements to have exit clauses. So they'll know when to back out of these projects if you know things are going south.
00:31:39
Speaker
um Yeah, so this is mostly for you know a long-term outlook for these data centers.
Co-Location of Data Centers and Nuclear Reactors
00:31:45
Speaker
So all of these deals are basically for data centers that are going to be constructed and coming online between 2035 to 2045.
00:31:53
Speaker
I have a question about location here, just in terms of the data centers. um Do we have a sense of where they're going to be built? And then second, um is the in terms of the location of the small modular reactors, um are those going is the reason that they have to be small is that they're going to be incorporated into more urban areas? Or what's the but's the importance of the size there in terms of getting them connected to the data centers?
00:32:22
Speaker
It's a good question. I think on the data center location side, there's a pretty strong incentive to be in areas that have existing other data centers and strong data connections. So Northern Virginia is a good example where yeah there's this really big sort of internet cable basically that kind of runs through Northern Virginia. um And there's a ton of data centers there. um And they have a ton of land to scale up.
00:32:51
Speaker
you They have some way, I mean, you you need to go out to like Frederick, right? Yeah. But there is land, you know, outside of Northern Virginia proper, right? um And the question is, you know, we've seen, will there be limitations on those prime areas, especially by 2035, right? You know, either because people don't want any more data centers in Northern Virginia, because all the good spots in Northern Virginia have been taken up,
00:33:21
Speaker
because they've got you no environmental issues. Maybe there is they're drawing a lot of groundwater. The aquifers are are looking rough. you know There's a lot of reasons why you could imagine these things being put into different places.
00:33:37
Speaker
I guess the question, Karthik, for you is, do you expect what's the level of co-location that you expect? like Do you think that they're going to build these in a completely separate place and like do some type of you know PPA kind of on paper situation?
00:33:53
Speaker
Um, or like, are they going to be in the same region? You know, grid connected maybe, but, um, kind of like the three mile island thing we talked about where it's, it's grid connected, but there's, you know, you're not actually running a cable from the nuclear power plant directly to the data center. Or is it actually going to be like really co-located in that way? what What are you expecting here as these things get built out?
00:34:16
Speaker
um As you pointed out, I think so far all these data center operators really like um existing infrastructure, which is why they want to go there. So if you know a data center already exists, there's great infrastructure. I'm just going to go continue building there. And as you pointed out, the challenge for co-location is then um why ah you know and only when you realize that a data center operator has to move to another geography where you don't have anything.
00:34:40
Speaker
Now, current data centers, they use a lot of water because you know they are air-chiller cooled, so they are water-cooled. um So assuming that data center operators continue the same trend, and the reason why I say that is because they tend to be very conservative. ah Data center operators don't like to spend a lot of money on something that they think may or may not work, so unless they don't have the surety, they won't go into a project. ah For example, with something like liquid cooling.
00:35:07
Speaker
or immersion cooling, which is you know going up a lot, um or at least is being talked about a lot. So co-location would make a lot of sense from an SMR standpoint as well, because you have a lot of water there and you also need water for like decay heat removal and things like that for the nuclear power plant. So you already are using water resources for the data center. So why would you have to search for a new spot with sufficient water for your nuclear power plant again, if your data center is already using it? So I think co-location would make more sense for the newer areas, for sure.
00:35:39
Speaker
And that's where I think the sector will move. um But again, the question is, why would anyone want to build a new data center and wait for the nuclear power plant to come online? If you can build a data center in two years and a nuclear power plant takes about, you know, five or six years to build the best case scenario, which is still a huge downtime.
00:35:59
Speaker
So is it just another big announcement to give the impression that they're making um progress in terms of moving towards clean energy for data centers? What'd you say? um I wouldn't say so. it's it's It's kind of like that. I think they are trying to put in tangible efforts to actually getting it ah realized. The only thing is the the the scale of the problem that they face is so high.
00:36:28
Speaker
that they have to be realistic about it. And their announcements also reflect that. They say, we are looking at early 2030s, 2035, 2039. Now, 5 gigawatt is definitely not possible by 2039, I don't think. If that happens, I'll be the happiest guy on this planet. I would definitely say that. ah Because I think we will need nuclear at some point in our grids to help us decarbonize. I think solar and wind alone won't cut it. But it's still a huge, huge problem to solve.
00:36:53
Speaker
ah And so I wouldn't call this green washing per se, but in the meantime, I feel these data center operators do not have the incentive to go low carbon in the short term, which worries me a little bit, um especially because there are no regulations or or any any policy that you know forces them to report emissions from scope two for electricity procurement.
00:37:18
Speaker
um We're seeing Karthik, you know, in other areas like e you know green hydrogen. we're We're beginning to see some weakness where companies have invested a lot of money. They're actually starting to build or break ground on some of these you know blue and green hydrogen production facilities. And they're not trying to find buyers for that hydrogen. And they're kind of struggling, right?
00:37:46
Speaker
um Do you think it'll be a similar type of situation where if you build, if you're one of the first people, you know, it's 2030, we're building these SMRs? People are beginning to break ground and you know, the costs are coming in higher than they expected. Things are more, the juice is more pricey. You know, the day center operators, are they going to stick to these commitments or is it going to be a situation where you're going to have a lot of nuclear capacity coming online? That's going to be struggling to, or some amount of nuclear capacity coming online that might struggle to find a buyer at the price they need.
00:38:22
Speaker
um I don't think at this point getting the buyer is the challenge. At least with the hydrogen side of things, I think the buyer has always been the major issue, right? No one wants to pay that much. It is a different dynamic there. Yeah. In the data center side, we are seeing people wanting to pay more for nuclear power. So they're acknowledging that nuclear is very expensive.
00:38:41
Speaker
But they want to pay for that premium. um They want to pay that premium for the reliability gains that they make. They believe that's higher than you know waiting to get grid interconnects and then getting unreliable power from the grid. um So I don't think that will happen. But again, you have to look at delays. You have to look at fuel procurement. ah but Lack of labor in the nuclear industry has always been a big issue. So you don't have enough skilled labor to construct. um So there will be delays, and it will get more and more expensive to build it.
00:39:10
Speaker
But the demand won't go away, at least as things stand, unless it becomes way too expensive for them to say, you know what, this is more than what we want to pay in terms of preview, which has happened in the past with NuScale. We have spoken about that on this podcast. So hopefully that does not happen, but it looks slightly better than the hydrogen side.
Conclusion and Upcoming Lux Research Events
00:39:27
Speaker
Well, I think we'll wrap it up there. Thanks so much for for tuning in. I want to remind you all listeners that We do have our forums coming up. We just had our forum in Amsterdam. These are in-person events. We have one in Silicon Valley coming up. We have one in Tokyo coming up if you're going to be in Tokyo. Do check those out. You can check out our other media. We have a blog. We have a lot of webinars. We do a lot of different publications.
00:39:54
Speaker
That's all on our website, www.luxresearchinc.com. And of course you can support the podcast by liking and subscribing, following us. We're on all your favorite podcasting apps. That really does help us out. Thanks again.
00:40:09
Speaker
Innovation Matters is a production of Lux Research, the leading sustainable innovation research and advisory firm. You can follow this podcast on Apple Music, Spotify, or wherever you get your podcasts. If you want more, check out www.luxresearchinc dot.com slash blog for all of the latest news, opinions, and articles. so