058 Trey Lauderdale, Founder & CEO of Atomic Canyon

Trey Lauderdale (00:00)
think of another part of the organization as the actual application of some of those models in specific components or specific.

verticals of the nuclear power arena. And that’s where we believe we can create tens of millions of dollars of value for enterprises.

We believe there’s a tremendous amount of efficiencies that can be created by leveraging artificial intelligence. So we have a very strong conviction that we’ll be able to build a sustainable company. And it’s a win -win -win. It’s a win for us. It’s a win for our customers. And it’s a win for society because as we help nuclear become better, faster, cheaper, we enable us to battle climate change, enable us, USA, to be energy independent from this technology that we’ve created.

So we think it’s great story overall. We’re super excited about it.

Mark Hinaman (01:56)
Okay, welcome to another episode of the Fire2Fission podcast where we talk about energy dense fuels and how they can better human lives. My name is Mark Hinaman and today I’m joined by Trey Lauderdale, the founder and CEO of Atomic Canyon. Trey, how you doing man?

Trey Lauderdale (02:11)
Hey Mark, good to hear from you. How are you doing today?

Mark Hinaman (02:14)
Excellent. I’m really excited for this conversation. I, you know, we were bantering a little bit before the call started. and I can tell we’re just going to get along. So Trey really excited, to hear about kind of atomic Canyon. You know, you’ve done a bunch of podcasts, already, which, which are awesome and disclosed,

what you guys are working on. for folks that aren’t podcast geeks like us, give us a little bit of background. Where’d you get your start?

Trey Lauderdale (02:41)
Yeah, so great to be on. Atomic Canyon is an artificial intelligence company.

that focuses on bringing next generation technologies to the existing nuclear power fleet. So from our view, nuclear power is the key to the future of energy. It is the best potential source for us to have clean energy, abundant energy. And in order for us to achieve that, we really need to drive efficiencies in the way we build new nuclear power plants, the way we help current nuclear power plants run more efficiently. We believe the entire kind of supply chain of nuclear

from uranium mining, milling, refining, enrichment, fuel fabrication, existing plants, new plants, recycling. Artificial intelligence is like the computer has been reinvented. So from our perspective, there needs to be very thoughtful and pragmatic ways to apply artificial intelligence across the supply chain. So our company has set out with a mission to help be that trusted vendor to help do this. And it’s been a ton of fun.

Seven months old, so we’re still, I’d say, early in our journey of being in nuclear, but this space is incredibly exciting. And it feels like every day there’s just a new announcement of whether it’s new laws passing supporting nuclear, like the Advance Act, bipartisan support. mean, in this day and age, what on earth has bipartisan support except for maybe nuclear power? So yeah, it’s really exciting. We love this industry and we think

Nuclear plus AI is just a great place to be in right now.

Mark Hinaman (04:18)
Yeah. Well, I, seven months as a software company, you guys are like sophomore status, right? I mean, as a, as a nuclear company, that’s like, that hasn’t existed since, know, like a hundred years ago, but.

Trey Lauderdale (04:25)
yeah.

Yeah, well, that is true. I think the point that you bring up is very well founded, which is being in AI, things move very quickly. New models are coming out on a weekly, monthly basis, especially in this whole cycle of large language models and generative AI. So you have this one sector that moves incredibly fast. And then in nuclear,

traditionally, as it should be, it is a very pragmatic, slow moving industry, because we need to be very safe and we need to be very thoughtful about the way that nuclear technology is deployed. So there definitely is this almost like intersection of really fast moving technology, plus a generally risk adverse industry. And as those two kind of come together,

There is a lot of, know, kind of just feeling like you’re getting turned back and forth.

Mark Hinaman (05:25)
Trey, why don’t you give us a little bit of background on you. How’d you get into this?

Trey Lauderdale (05:30)
Okay, well, I’ll give you the shortened abridged version because we only want to… Yeah, yeah. So, while we’re new to the nuclear space, my background, my co -founder and I, Christian, who’s an incredible AI technologist, we’ve done a lot of work in the healthcare IT or as it’s referred to now, digital health space. So back in 2008, I started a company Volt.

Mark Hinaman (05:35)
Yeah, you’ve told other people, but just for people to, yeah.

Trey Lauderdale (05:58)
It was the first company that brought iPhones into hospitals for clinical communication. And when people hear that, kind of get, you know, people turn their head and say like, well, what do you mean iPhones in hospitals? Like what else would people use? It’s hard to remember, but back in 2008, you know, the standard for communication devices, especially in the enterprise was Blackberry. Like everyone had Blackberries. And a lot of people would tell me, Trey, you’re crazy. The iPhone will never be in a hospital. My nurses and doctors would never

video calls or view the electronic medical record on a mobile device. And, you know, it was actually quite contrarian for us to say, embed on this mobile device or smartphone. To make a long story short, we were the leader in clinical communication. We put smartphones in the hands of hundreds of thousands of nurses and doctors, put a really robust infrastructure to let them communicate, receive mission critical alarms, did a lot of work with the FDA.

And it was really just an amazing experience to see a new technology, which was the smartphone, enter a kind of more risk adverse, slow moving market, which was healthcare. And again, healthcare should be risk adverse because if there’s a mistake in healthcare, someone could die. So that was a great experience, learned a tremendous amount about deploying technology to the enterprise, eventually sold that company.

to a public hospital bed company named Hillrom. We got to run the digital group at Hillrom, a lot of digital transformation to make Hillrom more of a bed company to more of a technology company. We then sold Hillrom to Baxter for $10 .5 billion. So a great transaction that was very successful. After that, did a lot of angel investing, kind of fractional executive work.

some work with venture capital and PE, all in digital health for a couple of years. And those past few years, it was all about artificial intelligence and healthcare I’d say is definitely on more of the leading edge of applications of artificial intelligence. There’s been a lot of work in digitizing health records, which has enabled kind of healthcare in general to apply AI in a lot of really advanced ways. So examples are helping a company that was supporting radiologists by enabling computer vision that helped those

maybe see cancer or other things that the human eye might miss. One of my favorite companies was leveraging automatic speech recognition to record a physician patient encounter and automatically create that note for the doctor in the electronic medical record in a matter of minutes with like crazy accuracy, like 99 % accuracy across 25 languages. So for me, that’s what I really got to realize that artificial intelligence has gone from the R &D

phase where it’s kind of a cool project and PhDs are working on the science to no, no, we are ready to apply this in again, a risk adverse industry, which is healthcare and really see material impact. I mean, there were physicians that would tell the company I used to work

this artificial intelligence application has allowed me to stay in medicine. I’m getting two hours back a day. I can go home and spend time with my family. And healthcare means a lot. Like all of us have to engage in the healthcare system eventually. So we want applications like this out there. And you get to see that, know, as AI is real, like it is going to have substantial impact upon humanity. So got to see a lot of firsthand AI applications. And you might say, okay, Trey, well, get to nuclear. Like, how’d you get in nuclear?

Three years ago, my wife and I, moved to San Luis Obispo and the house I’m in right now, we’re actually 10 miles from a nuclear power plant, which is Diablo Canyon nuclear power plant. So if you live in the proximity of a nuclear power plant,

and you’re in a smaller community, which nuclear power plants tend to be in more smaller rural communities. They are a key foundational partner of the ecosystem. So people I’d meet everywhere work at the nuclear power plant engineers. of my favorite is my son’s basketball coach was an engineering manager at Diablo Canyon. And I’m generally a curious person. So I’d ask people, know, nuclear sounds really interesting. Tell me about this. And they’d walk through the science of nuclear.

And it really is a modern day miracle. mean, the fact that we can split atoms and through that create heat and we can use that heat to create clean energy and the by -product just sits in canisters. Like it’s right there and we can actually recycle that by -product, no carbon emission. It is a modern day miracle. So I was just incredibly fascinated to learn more about this incredible technology. But when I’d ask, so tell me about your IT systems and some of those computer programs that

using they would tend to laugh and almost kind of be like wow well let us show you what we got here and you quickly realize nuclear power is this incredible technology but there’s been a tremendous amount of regulation

there’s not necessarily the most advanced IT systems that are being used. So that’s when it really dawned upon me that there is an opportunity here. And I started researching and seeing that the Biden administration has been very pro -nuclear. The Department of Energy has across the board put tons of initiatives in place to really reinvigorate our nuclear fleet. And there’s almost this like excitement from multiple fronts where you’ve got on one side, it’s national security, China and Russia and some of our foreign adversaries.

are expanding and building nuclear and we as Americans need to make sure that we are the leaders in nuclear technology. And separately there’s the climate change opportunity where look, climate change, we’re starting to see the impacts of that.

So you have this just bipartisan excitement and support for nuclear. So as I was seeing that and kind of pulling one and one together, it really dawned upon us that artificial intelligence has a key role to play in enabling and supporting the nuclear power expansion across the United States. So it’s easy to come up with ideas and talk about, you know, concepts. I’m a big fan as an entrepreneur, you have to take action and you could sit here and think and ponder and pontificate all day.

What differentiates, you know, just research from actual application is building. So I told my friend Christian, who eventually became the co -founder, we’re going to get into nuclear. And this was around Thanksgiving. We need a data set. Let’s go find one. So we kind of scoured the internet, did as much research as we could. We found the nuclear regulatory commissions, Adams database, a phenomenal source of open publicly available data.

We downloaded all of those documents, all 52 million pages. We then put together a series of kind of computer vision, optical character recognition, kind of fancy AI ways of saying we understood what was in all those documents. And then we launched a product, Neutron, totally free. We launched this in March, which is a way that enables any nuclear power plant.

anyone who’s in the nuclear industry to search and find information on atoms in a more efficient format. And the problem to be solved here is the nuclear regulatory commission in this atoms database, the interface that they have is a little bit cumbersome today to try to find information. So we want to do leverage AI to make that really easy to search and find. yeah, well, I love my friends on the NRC. what I’ll say is there’s opportunity for improvement.

Mark Hinaman (13:21)
That’s an understatement. Cumbersome’s putting it plainly.

Trey Lauderdale (13:32)
And I give the NRC a lot of credit because they digitized all these records. Like, so we could not build our AI if they hadn’t gone through the process of taking the microfilm and putting it online. Look, it’s a nuclear, it is a team effort. We all got to be working together. Yeah.

Mark Hinaman (13:44)
And it’s pretty early on that they did that, right? mean, some of these documents are like from the sixties and fifties, like, yeah, they’re digitization.

Trey Lauderdale (13:52)
no, no, my friend, my co -founder Christian, he loves this because if you think about really good engineers.

What do they love? love big, challenging, meaningful projects. So being able to get these documents that are from like the sixties or seventies that are like typewriter documents with all of these acronyms and be able to have computer vision, actually understand what’s in them. Huge technology challenge. And Christian likes to share with me that the breakthroughs in AI and computer vision and everything else, what we’re doing, we couldn’t have done a year ago. That’s how fast

is moving. So it’s super, super exciting. So we launched this product, Neutron, totally free. We’ve got tons of users on it. We’re getting lots of feedback. The nuclear industry is some of the nicest people I’ve ever met in my entire life. They have been so welcoming across the board. We’ve made a ton of contacts and a lot of relationships. from

What we realized as we were building AI and nuclear is nuclear is really complicated and the words that are used and the acronyms are really challenging for AI models to understand. So in order to solve this, everyone comes to us and says, Trey, build us a chat GPT for nuclear.

And what we’ve explained to them is there’s actually a missing piece of the puzzle here, which is you can’t just go and use an open source LLM and build a chat GPT for nuclear. It hallucinates too much. And you hear the word hallucinate. That’s us AI guys fancy way of saying it makes stuff up. So if you’ve ever used like a chat GPT or any of the large language models.

You don’t know if what it’s saying is true or not. Like lot of times it’s accurate and it’s getting better and better. But in many cases, it’ll just like make up stuff that’s wrong. So that is very bad in nuclear. cannot hallucinate. You cannot make things up. So it has solved this problem. We realized that we actually needed to train the artificial intelligence to understand nuclear terminology. So to solve this problem, you need a supercomputer, you need a really powerful computer.

So we went and we got connected with the Department of Energy, the Oak Ridge National Laboratory and Frontier, which is the world’s fastest supercomputer. It just so happened their team was looking to solve this same problem. So we were quickly allocated time on Frontier to use their computer to build out and train what’s referred to as sentence embedding models. So this is actually training artificial intelligence to understand nuclear terminology. And we’re actually going to open source this work because we

the entire industry for free should have access to this technology because we think there’s going to need to be an army of AI developers building applications. And we want to really start to set the foundation and put foundational technologies in place to enable.

know, many, players to enter this space. So that’s been a lot of the work that we’re doing now. It’s a lot of learning. It’s a lot of putting foundational AI components in place to really invite and try to pull many, many more AI developers into the nuclear space. Because I think this is going to be one of the most important applications of AI is going to be to help nuclear power.

become more efficient, more cost effective, and ideally the preferred technology to help clean energy. So that was a lot. So let me know where you want to take it from there.

Mark Hinaman (17:17)
It’s all good. You’ve given the spiel before I can tell, but that’s okay. It was helpful.

Trey Lauderdale (17:21)
Yeah, yeah, yeah. I get excited about it. I love what we’re doing. I could not pick another, a better industry to be in right

Mark Hinaman (17:28)
Yeah. why this supercomputer? Like, why do you need that versus training on like a different model? And I imagine that like helps set you guys apart a little bit, right? Being one of the first movers to this place. Like anytime you got that, creates a little bit of a moat, right?

Trey Lauderdale (17:41)
Yeah, so when you look

Yeah. So when you look at artificial intelligence companies, you’ll notice they tend to raise a lot of capital. Like they have to go and raise 10, 20, $30 million. And a lot of the rationale behind that large raise is to train specific models. You need a tremendous amount of GPUs, graphical processing units. So you need to go and purchase multiple Nvidia H100s, or you need to partner with like a Microsoft or a Google.

And that’s expensive. Whether you’re, if you raise the money as a VC backed company, you’ve given up a ton of equity to go buy a bunch of computers. So from our perspective, the ability to leverage government resources and have a partnership with Oak Ridge National Laboratory, it really helps us, you know, twofold. Number one, we don’t need to go raise the 10 million plus dollars to build out these kinds of clusters of supercomputers. So from a company perspective, and that enables us to

more equity and more control over our future and our destiny. And then separately, Oak Ridge, as I’m sure you know, is one of the nation’s,

premier research organizations. It’s part of the Department of Energy’s national laboratories. That group does some of the most advanced work on the planet in both nuclear and artificial intelligence. beyond enabling us to have access to resources, that would have been very, very expensive for us from an equity and venture capital perspective. We as a company get to husband our resources.

extremely capital efficient because of this partnership. then separately, we get the almost halo effect of, know, Oak Ridge has hundreds, if not thousands of companies that have come and want to work with them. The fact that they have selected us as a partner is a real testament.

part honestly to our co – to my co -founder Christian, he is a technological genius. And if he was here, he’d be blushing, but I always say he is a phenomenal genius. He’s one of the best people I’ve ever met in artificial intelligence. And the fact that they’ve elected to work with him and our team is a testament. We are not, you know, what some people say is like a rapper on top of open AI where

There’s nothing wrong with that, which is you’re more of just the application itself of the artificial intelligence. We are a hardcore AI research and development shop. We roll up our sleeves and we build very advanced AI technology. So while we might not move as fast as some of the other kinds of AI companies that are out

What you will find is we are very pragmatic. We pick specific problems, we execute against those problems, and we solve real world applications of AI in nuclear. And this is a lot of the playbook that we used in healthcare, where you could come and you could say the smartphone, you know, out the gates could solve all these problems and you can integrate all these advanced alarms. But really you need to figure out a pragmatic methodical methodology to introduce new technology in a

kind of risk conscious way, which is how you gain traction in a industry that is, know, life change or life critical and has the potential to have really positive or negative impacts. You have to be very, very smart in how you will deploy this technology. So a lot of what we’re doing is leveraging the supercomputer to build artificial intelligence, which helps people use AI to just find documents. Like if you think about AI,

It can be scary. Like sometimes I tell people I’m an AI in nuclear and they’ll say, true, that that’s the theme of Terminator. What are you doing? That sounds insane. So I have to really pull people back and say, well, the first thing we’re going to do is use AI to help you find documents. You’re like, okay, that sounds pretty, you’re risk adverse. And then we prove ourselves. We prove that AI is reliable.

It doesn’t hallucinate when it’s built by us. It works really well. And you earn that foundational level of trust, which then enables you to tackle the more complicated, robust or riskier problems. So it’s a very kind of stepwise process of applying new, new generation technology to kind of a risk adverse industry. So.

Mark Hinaman (22:04)
Trey, and so you talked about open sourcing this tool and being capital efficient. mean, I think that’s brilliant, right? But tools also free right now. So while you guys are building the service and demonstrating, capability, I mean, that’s, intelligent and also really helpful for, folks. I’ll say I really enjoy working with small software companies, as a user because it’s more common that they’re responsive. They give rapid feedback. They’re willing to help you solve problems.

as you use their tool and I mean as a user of the Neutron tool that’s been my experience certainly.

How does it scale? What does it look like? How are you guys open sourcing this and what will that continue forever? Or is there, I guess what’s the kind of longer term

Trey Lauderdale (22:51)
So from a go -to -market perspective, they’re the first product we launched Neutron absolutely free. And a big part of that is we come at nuclear with a tremendous amount of humility. And we believe as is important in nuclear transparency is key. So a key part of our mission vision values is about transparency and honesty, because I think it is going to be incredibly important when dealing in nuclear power.

that you are as transparent and open as possible. So when we launched Neutron, we’re very clear with the constituents that we engage with that we’re new to nuclear. We’re here to learn and we think we’re fast learners. And I think we made a ton of progress, but we still know that we’re the new kids on the block in this space. So Neutron was our way of number one, first giving back to the community and starting to demonstrate that artificial intelligence can be used in nuclear. Separate from that, from an open source perspective, the first models that were

building, we believe transparency is very important in artificial intelligence, especially when it comes to nuclear power. So we plan on open sourcing on hugging face, the, the, models themselves, the training data, all the information. So people can go and audit and ideally improve our models. That’s the whole spirit of open sources. You put the code out there, people use it and you all together as a community improve the overall data set. And believe it or not, I like to say.

the Nuclear Regulatory Commission, whether this was intentional or not, the ethos and the spirit of open source is embedded within that organization. So if you think about the NRC and the Adams database, what is it? It is every nuclear power plant throughout the United States sharing information and sharing data publicly that’s available to everyone else. So that is the exact same concept of open source technologies. So from our perspective, these first foundational models, we believe it’s really important

We share those with the industry, we get feedback, and we believe the rising tide helps all boats float.

from a go -to -market and monetization perspective. Yes, we will eventually have to make money to grow the company, to raise capital, and to build a sustainable organization that can have the impact that we want to have across the nuclear industry and across society. So to enable that, we’ve had a number of discussions. We’re not ready to announce anything, but there’s a number of not only nuclear power plants, small modular reactors, the existing large reactor OEMs,

consulting companies, you name it, that we’ve been engaged with, where we’re taking parts of Neutron, what we built that’s publicly available, we can deploy that in more of an enterprise format, whether that’s on -premise or in a private cloud. And then there’s applications that we’re building on top of our Neutron platform that help create value for specific constituents and stakeholders in the nuclear power arena. And those are the products and services we will be charging for. So you can think of

Neutron and the sentence embedding models as kind of the platform that we’re building. We’re open sourcing, you know, big components of that. And we’ll continue to keep the methodology and the spirit of open source and a lot of our foundational models that we’re building. But then separately think of another part of the organization as the actual application of some of those models in specific components or specific.

verticals of the nuclear power arena. And that’s where we believe we can create tens of millions of dollars of value for enterprises. And we then cost share in some of those savings by helping organizations run more efficiently. We can then charge money and create a sustainable business model. So that’s really been the go to market. And yeah, it’s been super exciting. This industry has been very, very receptive.

We believe there’s a tremendous amount of efficiencies that can be created by leveraging artificial intelligence. So we have a very strong conviction that we’ll be able to build a sustainable company. And it’s a win -win -win. It’s a win for us. It’s a win for our customers. And it’s a win for society because as we help nuclear become better, faster, cheaper, we enable us to battle climate change, enable us, USA, to be energy independent from this technology that we’ve created.

So we think it’s great story overall. We’re super excited about it.

Mark Hinaman (27:18)
Yeah. You mentioned early on that you saw some of the IT systems in the nuclear plants and they perhaps weren’t as far along as some of the systems that you were used to implementing in hospitals and in healthcare. Were there like immediate parallels that you saw? Like, do you have an example of something that comes to mind that you’re like, wow, we already did this in healthcare. Like we could easily upgrade this for

Trey Lauderdale (27:42)
Yeah, yeah, so.

Yeah, so I mean, well, not that we’re starting here, but I chuckle because I used to tell people pagers, know, like the 1980 pagers, I used to tell people hospitals and doctors are the only ones that have pagers. Like that’s it. Like that’s the last industry. And lo and behold, I found nuclear where they still use pagers. it’s kind of, me and my friends are kind of, we kind of get a kick out of it. We’re like,

Mark Hinaman (28:08)
You’re fandoming in the nuclear power plants. Like

Trey Lauderdale (28:13)
our war against the pagers never seems to end. And here we found another industry that’s got pagers. So we’re not starting there. But I think it’s definitely an indicator of an industry that might need some digital transformation is when you run into pagers. that was it. We got a hoot out of that. But separately, where I’d kind of place

nuclear power is if you flash back to healthcare. So healthcare, the real foundational technology that has really kind of enabled this whole healthcare IT digital transformation is the electronic medical record. So the

really didn’t gain a tremendous amount of traction until there was the Meaningful Use Act, which happened in 2008, 2009, which was part of the American Recovery and Reinvestment

which required all hospitals and all physician offices to transition to electronic medical records. The first, they use the carrot. If you did that, could get, you’d actually get paid out from the government. So there was government subsidies to move you to the EMR. If you didn’t do it after a few years, there was the stick. So carrot and stick. We actually got fined if you did not move to electronic medical records. So that became a big catalyst to enable Epic, Cerner, all of these electronic medical record companies to really become pervasive and enable the

transformation of the healthcare industry. And you have people say on both sides, some people say, God, that was brutal. Like we shouldn’t have done that. It caused us to like deploy these systems that, you know, we’re not necessarily the best user interface. Other people have said it’s been a huge success. I think, know, the truth is always somewhere in the middle where it’s good that we got the digital transformation. Maybe it could have been a different path to get there, but as a result, you had this digital transformation. when the AI wave hit.

there was all this foundational data that enabled artificial intelligence to get hold of tremendous amounts of venture capital investment and private equity investment to really drive that industry forward. So I’d say nuclear is kind of in a space where when you look, there is no equivalent of the electronic medical record.

or an operating system in the nuclear power space. There tends to be a number of what I’d say are like traditional ERP, Enterprise Resource Planning Services like SAP and these other like big clunky systems that have been kind of customized and you get like these big consultants that you pay way too much money to like customize these

I don’t want to say legacy, but it’s a lot of 20 -year -old products that are there. It’s like you’re a perfect example of enterprise, slow -moving products that really weren’t built with the end user in mind. So when I see that, I believe there’s tremendous opportunity across the entire supply chain. So I

I love the idea of starting with the current nuclear power fleet, like going in and saying, Hey, we’re going to leapfrog and we’re going to bring the latest and most advanced technology and artificial intelligence. We’re going to find these little crevices of where AI can be applied to just supercharge your organization. And we’re going to do this as we build a platform that enables multiple modules of AI to be built and make the current nuclear power fleet super effective. But separately, when you look at like the construction of next generation reactors and the whole

process to help get these reactors approved, get them built. There is going to need to be almost like a construction nuclear operating system as well that’s built on AI. So I think there’s huge potential because in nuclear, everything is documented. And here’s kind of like what I think is the key insight that I don’t think everyone realizes yet.

The fact that this industry has been forced to document everything means we have the best data source. I might say on the planet as an industry, nuclear is primed with tremendous amounts of data. What does AI need? You need a tremendous amount of data. So I don’t think people realize yet, but artificial intelligence is going to absolutely transform nuclear power. Like it is going to enable this industry to move way faster

ever expected. And it is my prediction in the next five to 10 years, nuclear power as a segment will be the most advanced from an AI perspective. We have the data, we have the resources, and now that we suddenly are seeing the need, mean, ironically, a lot of the energy is coming from AI itself. So you have this world where, you know, Microsoft is saying we need to build a five gigawatt data center. It’s like, holy, holy crap, like five gigawatt data center. Like that’s like five

P1000s plugged into a data center. Like that is insane. So you got like this huge need of power coming from

AI, electric vehicles. So the fact that that power is going to then force us to figure out how to make nuclear more efficient, AI is the pathway to make nuclear more efficient to then enable nuclear to help the data center. So it’s almost like this really ironic circular relationship of nuclear needs AI to become more efficient, but AI is going to need nuclear if we’re going to keep that exponential growth curve of training these large language models and these other

very energy consumptive kind of AI resources. So it’s a super exciting time. Like this little intersection, I didn’t realize when I started the company how exciting it would be, but it’s a blast. We love it.

Mark Hinaman (33:54)
Yeah. so I was just listening to some today talking about data centers and how much of the industry, the existing industry data centers are occupied with like AI versus just classic data center stuff, right? And like what data centers have enabled in the world. And like the example that was given on the podcast that I really liked was,

Uber or ride hailing services did not exist and weren’t even really conceptualized. mean, they kind of were, but they kind of weren’t. And I think when people, when we have technological, you can say revolutions or step changes in technology like AI, I think it’s difficult for people to wrap their minds around some of the like real benefits. So you’ve given the example, I mean, you guys have your platform and base that you’re using right off the bat,

What does this look like for people within the nuclear industry? Can you give maybe another example of like the potential here and how it might impact in the future? Like what does the workflow look like that could expedite and create some this value and allow people to do their jobs faster or to have nuclear power plants operate more lean?

Trey Lauderdale (34:57)
Yeah, yeah.

Yeah, yeah. So let’s kind of blue sky this out a little bit. And again, when I say statements like this, I like to preface it by saying,

I remember when I used to go in 2008 and tell nurses and doctors and hospital administrators, one day you are going to have an iPhone in the hands of each of your clinicians. And in that iPhone, a doctor will receive a notification about a patient and they’ll be able to hit a button and view that patient’s medical record on the phone. They’ll be able to view the real time physiological monitor.

in an FDA approved format. So you will have streaming vitals near real time on a phone. And then you’ll be able to tap a button and video call in with that patient who will have their own iPad or similar device in front of them. And you’ll be able to see the patient in real time. Like I say that today, everyone’s like, well, duh, like, of course we could do that. Like what else would we do? Like that’s very simple. Like that’s expected everywhere.

In 2008, that sounded like witchcraft. It’s like, Trey, you’re insane. Like, what are you out there smoking? There is no way that would ever possibly happen. yeah, well, know, blackberry, that’s one thing, but yeah, but people, like, you just couldn’t even conceptualize how these devices, these mobile computers in your hand, how they are going to transform workflows. Like it’s just, you just couldn’t conceptualize this because you never had the technology before. So when you look at artificial intelligence,

Mark Hinaman (36:11)
Yeah, on my Blackberry, that’s never gonna happen. Yeah.

Trey Lauderdale (36:35)
There’s a couple of things I need to say, because this is a different type of technological transformation. So the first thing is the technology is moving way faster than the smartphone.

So I’d say when like the iPhone 1 came out, like it was cool. Don’t get me wrong, it was a really cool device, but it wasn’t really until like iPhone 4 or 5, 6 that it was just like unbelievable. Like you had this app ecosystem and they really figured out like the user experience and they like, like everything like it took like four or five years for like this to really be like an amazing device.

With AI, you’ve seen the four to five year of technological improvements condensed into four or five months. AI is changing really, really quickly. So that’s important, number one, to recognize. If AI stopped today,

I could do everything I’m about to describe to you. But AI is not going to stop. It keeps getting better and better. So that’s just one thing to keep in mind is we’re going to be in this very challenging environment where the technology is going to be way more advanced than our ability to consume that as a risk -inverse enterprise.

So that’s where companies like Atomic Canyon or come in is where all this like really advanced stuff is happening. How do we take it, make it reliable, and then actually apply it in a way that creates value for the enterprise. So let’s give a couple of examples. So, you know, today you have a nuclear power plant that’s operating and they’ll receive a request from information from the NRC. Maybe there’s questions about

the backup generators and the processes they’re using and the NRC will send a request for information. So what happens today is that at that nuclear power plant, there’s a team of regulatory people that get that request and say, okay, let’s go first search all the NRC’s website and try to find an Adams, all the relevant information. It’s almost like case law of how do we understand how others have responded to the NRC so we can get our response ready.

It takes a couple of weeks of, you know, a half dozen people searching NRC. Then internally, it’s like, okay, let’s go find all of our information that we’re going to use to respond to that request. Then you got another team of people looking through their record management and all these documents and figuring, oh, we can’t find this piece of information. And then they’ll find that one person in engineering that’s been there like 30 years that knows where everything is. And that person will get in and they’ll find it. And so they then gather all this information and they put together.

a draft of what they’re going to send back and they review it and they bring in consultants and those consultants review it and then you know three to six months later they send their response back and then it goes back and forth between NRC and the nuclear power plant. So within the next few years the way this is going to

is the NRC is going to send a request in that is going to get caught by the AI at the nuclear power plant. That AI on behalf of the individuals will search the entire NRC website, will find all the information on atoms of everything that’s relevant, will gather all that information to help create the best potential template of how the response should look. Separately, there will be a separate AI agent that lives securely inside the enterprise behind the

wall of the nuclear power plant. And that AI agent will go search all their internal data, will understand what’s the relevant data to place in that document. And the AI will create a first round draft. And it will point to all the documents that it recommended and that it pulled from to create this first round draft. So then the employees of the nuclear power plant won’t spend their time hunting and searching and looking for data. They will get to spend their time doing

what humans are way better at than AI, which is thinking critically and understanding how do we want to position this? What’s strategy of how we position this? What’s the right information to share? So it enables humans to do what humans do best, which is think creatively and actually be thoughtful in their day to day versus searching and hunting and hiring consultants to go find data. So that world is coming, whether it’s us at Atomic Canyon or others, that will be the way the world looks in the next few years. And the reason nuclear is so well

to do that is because we have this tremendous data set to train the AI models off of. So the idea of consolidating and collapsing these regulatory processes from months or years to like days and weeks, that sounds, when I tell people, farfetched. It sounds like something that, it’s the equivalent of me telling people your iPhone will view FDA approved waveforms as a doctor.

Like when I share like that, this is going to happen in AI, like the exact similar parallel is there. You just have to be very thoughtful in how you integrate different components of this AI to find your way to that kind of true north of AI automation. So that’s one example. I can give others of your day -to -day work of what construction workers will be doing when we’re building next generation nuclear power plants or your current existing operators and maintenance crew.

the idea of waking up in the morning, getting to work, putting on your Apple Vision Pro 6s, which will kind of look like sunglasses, and then going in and those Apple Vision Pro 6s will have cameras built into it. So they will see everything that the person is seeing. And as that person goes through their day -to -day work and they’re looking at a pump and they’re doing their work order to change different components of that pump.

their computer vision will be telling them, move this screw to the right, press this button. the light turned red. That means you need to go through process ABC. And then there’ll be a microphone and speaker that will be telling them step by step what it is there to do. And all of that will be powered by artificial intelligence. It’ll label the operation and the maintenance groups and the construction workers to move what, three, four, five times more efficiently. So when you think about

the modalities of AI, a lot of what we’ve been talking about has been how do you figure out how to build like a chat GPT for nuclear? How do you streamline regulatory processes? From our perspective.

this nuclear operating system, the way we’ve designed it and built it is it’s going to be multimodal. So we’re going to start with the most risk adverse way of tackling AI and nuclear, which is all regulatory, is taking and finding information. But our vision is a nuclear operating system, which enables multiple modalities of AI, including computer vision, automatic speech recognition. And I think that’s part of the value that we bring is we’ve

on all of these different types of AI in our previous lives. So understanding how AI can be deployed across multiple modalities is a skillset that we bring to the industry, which we think is really unique. And I think it is going to be an incredible world. It is gonna take time to get there. We’re not just gonna jump to this magically, but that’s what we’re excited about is, you you go and you figure out what’s the problem.

How do we solve that problem with this new AI technology? And how do we make sure it drives value? That daisy chain of events is what we’re really, really good at. hopefully that gives you a couple ideas of where we’re taking this.

Mark Hinaman (43:55)
Now, I love that example. I think it’s brilliant. And it really feels like the future. I’ll comment, right? Well, you gave like a knowledge worker example, and then you gave like a real life tangible, like, okay, well, here’s a trade skilled worker that will actually benefit from technology advancement. it’s easy. I mean, as a technocrat, it’s easy for me to see this progression.

Trey Lauderdale (44:04)
yeah, future’s gonna be here faster than you know. It’s gonna be, yeah, it’s gonna be exciting.

Mark Hinaman (44:23)
But I think people underestimate like how the progression happens. it’s, it’s like a continuous improvement. And like you said, it’s, focusing like innovation doesn’t just jump to like robots doing the job and doing the welding on site, but it’s one step at a time. You solve a small problem and then, okay, great. That’s on the books. Then you go and solve the bigger problem and more people that are solving those problems along the way. And then sharing that knowledge, like that’s how you get there. So I think there’s awesome potential.

Trey Lauderdale (44:50)
Yeah, well, like, yeah, the parallel I love to give in healthcare. And again, I draw a lot from that because it’s 15 years of my experience, but I think it’s super, super relevant was a mistake we made in healthcare.

Mark Hinaman (45:01)
I think it’s helpful to bring that outside experience to other industries. I think it compounds and it’s really productive.

Trey Lauderdale (45:08)
An example I love to give is when we first brought iPhones into healthcare, we were naive. I was 25 years old, hadn’t done this before. So I used to go tell nurses and doctors, Hey, we’re going to integrate all your mission critical alarms. We’re going to enable text messaging. We’re do all this out the gates. And the doctors and nurses would look at me and be like, are you crazy? We are not getting rid of this code blue pager. like, when this pager goes off, we have to go save a life. There is no way we are trusting your smartphone with this pagers notification. And I was like, Ooh,

Okay, lesson learned. So we pivot and you very quickly realize.

The first foundational step is getting smartphones in the hands of nurses and doctors and then enabling them just to send text messages, like literally a directory and you can send someone a non -critical text message. And lo and behold, they got huge value from that. And we got to learn, we got to realize how do we make this reliable? What happens if the iPhone loses wifi? How do you have a backup plan? How do you have escalations for messages? Like you learn all this information and then you improve that foundational tech.

And then people gain confidence that the technology will work. And then from there…

After we show that text messages work, we built this directory. We then said, okay, we are now going to send that code blue notification to a smartphone and you’re still going to keep your pager. Like we’re not going to take away your pager. You’re just going to get it twice. And that’s how we’re going to go live. And then when you go live with that, and then the first code blue that happens, everyone’s smartphone like goes off in less than a second. You’re like, oh, code blue, like a customized ring tone and everything. And then like 30 seconds later, you hear the pagers and then.

That happens a couple of times, and then people quickly realize, you see a drawer full of pagers, because people have then gained trust that the smartphone works, and you built all these steps of extra reliability where if someone doesn’t respond, it escalates to a backup caregiver. So a perfect example, when we start deploying AI nuclear, we’re not going to say, the AI runs your nuclear power plant. That would be insane to go there. You start by saying, the AI will gather your data for

search and find, and then you show that it works. And then once you’ve built that reliability, you build the next level of trust, then you build the next level of trust. So it’s a journey. It’s a journey across all the modalities. yeah, we’re super excited. We got some phenomenal partners that we’ll be announcing in the next few months of people that we’re working with. I think people are going to be really excited when they hear some of the progression that we’ve been working on the last few months.

Mark Hinaman (47:39)
Trey, this has been great. We’re coming up on our time.

got two more questions for you Trey. One of the things on our list was that every tech company might become an energy company ultimately, right? Like this is where the nexus is going to end up. What are your thoughts on that?

Trey Lauderdale (47:56)
Yeah, so when you think about big tech, when you think about like Nvidia, Microsoft, Google, I think we as the nuclear industry are going to have an awakening that is going to occur, which is the magnitude of the size of some of these big tech companies is just like, it’s very hard to comprehend.

So let me give an example. Our good friends had a… Oh, yeah. No, no, no. It’s like… Oh, no, no. Here’s my favorite example. So, NVIDIA. So if you know NVIDIA, they are the producers of the GPUs that almost everyone uses in AI. So NVIDIA, they’re the ones, like their stock has taken off. They were the world’s most valuable company for a short period of time. The other day, their stock gained 1 % in value. So the stock went up 1%.

Mark Hinaman (48:26)
It’s like country -sized companies, right? Like there’s more.

Trey Lauderdale (48:53)
The equivalent of that is, you know, is roughly $75 billion. $75 billion in market capitalization gain with 1 % of NVIDIA. The largest provider of nuclear power in the United States is Constellation. Constellation has a wind fleet, they’ve got clean energy, and they also have, I think, somewhere like 24 nuclear reactors. The market cap of Constellation, the last time I checked, is roughly $70 billion. So, like, let that sink in.

a 1 % 1 % shift in Nvidia

is the equivalent of constellation energy. like we as a nuclear industry, don’t think really understand as AI, as Microsoft, as Google, as Nvidia, as Meta, as these gargantuan trillion dollar plus companies continue to grow. Like what’s their growth strategy? All of them are investing tens of billions of dollars in artificial intelligence because they see this is

next wave of growth and they are they are all fighting to capture that next wave. So what has been the bottleneck to that wave? It has been access to GPUs. The supply chain was not ready to produce enough GPUs to keep up with the growth. That’s starting to work itself out. You got other GPU providers coming in, Microsoft and Amazon is everyone. They’re all getting into GPUs themselves. They have thrown money to fix that problem because when the big tech machine moves,

There’s very few things I can get in its way. The next bottleneck, which we’re already starting to see, is going to be energy. When you want to go build a 2, 3, 4 gigawatt data center, where on earth are you going to get that power from? So as that becomes the leading bottleneck, I think we are going to see actions from big tech that blow our mind. And it’s funny. You see big tech making statements like, oh, well, we’re not going to get into.

you know, financing nuclear and that, I won’t say which companies have said that, but a couple of big tech companies have put a little bit of arms distance between them and nuclear. And it’s so funny. Like if you ever see NFL coaches or college football, I’m a big college football fan and you have a college football coach who’s like, there’s absolutely no way I am going to coach Ohio state or I’m going to leave. I am so happy where I’m at and that, and then like two days later, it’s like, I’m so thrilled to be announced as the head coach of your Ohio state or

the Florida Gators. Like, it’s like, you know, they’re not going to come out and say

But look, when big tech has determined that they need energy and that that is starting to affect their value of their company and their growth trajectory, there is going to be an unloading of tremendous amounts of capital. How does that unfold? I do not know, but my instinct is there is going to be some moves that happen in the nuclear power space that are just going to be tremendous. And these companies are all going to figure out in the next three to five years how they become energy companies.

because energy will be the bottleneck, whether that’s through joint ventures, whether that’s through straight out acquisitions. These companies will solve this energy problem.

because they have the capital to do it. And it is going to be the problem that’s in their way. And I think AI is going to play a big part of that. So we’ll see how that all plays out. that’s my thesis is all the big tech companies within five years, they will all be energy companies, probably nuclear energy companies, but we’ll see how it all plays out.

Mark Hinaman (52:26)
Well, we’re lucky that we’re recording this so we can actually, you know, hold you accountable on that prediction. Right.

Trey Lauderdale (52:32)
yeah, no, that’ll be fun. It’ll be good, it’ll be good.

Mark Hinaman (52:35)
Last question for you, Trey. How can people help? Looking forward, how do people help move and advance this to bring the industry

Trey Lauderdale (52:47)
So I’d say that when you look at the nuclear power space, we need more people. there is just the existing. Yeah, we need more. Everyone come on in. We need everyone to come this way. So when I look at this, again, drawing a parallel.

Mark Hinaman (52:56)
Literally just human capital. Like we need more people working on these problems.

Trey Lauderdale (53:07)
to healthcare IT, my previous industry. I remember in 2008 when I used to go try to get VC fundraising, because I didn’t have the resources to sell fund back then. When I used to go speak to venture capitalists and others, it was so funny. They all would say, oh, we don’t do healthcare IT. We don’t think there’s money to be made there. It’s not a great opportunity. All these fancy investors on Sand Hill Road and Silicon Valley. And lo and behold, 10 years later,

Single one of them to the tee has a digital health fund or a digital health thesis Like it is so funny how things have switched to the complete 180

In nuclear, like nuclear is going to be the key to unlocking clean energy is going to be the key to keeping artificial intelligence moving forward. It is going to be the key to energy abundance of our society. So we need number one, the entrepreneurs to enter this space. need the technologists to enter this space. need the funding to enter this space and it’ll all come like this will all unfold as the market opportunity presents itself. And eventually you’ll see one after the next, you know, VC launched their nuclear

fund or they’ll figure out some cute name for it like their atomic energy or their vision fund. There’ll be all sorts of that. But we need the entrepreneurs. need more of the entrepreneurs entering this space. I, as a nuclear entrepreneur, will tell you I’ll be the first one to field calls from new entrepreneurs coming in, giving you advice, helping you make connections. This is a warm and welcoming

environment for those coming in. I think the current people that are in nuclear state as long as you can, the operators of the fleet, hold off your retirement a few more years. We need young people entering the plants from a construction standpoint.

We’re finishing Vogel three and four. As we look at new companies or new organizations that are building next generation reactors, we’re going to need people in construction, supply chain. We as a society, you just need more people to enter nuclear. There’s plenty of jobs. There’s going be a growing space. So I’d say come in

I know it’s scary leaving an industry and going into a new industry. I just personally did it. My investors who I had, who some of them put money in Atomic Canyon, some didn’t. When I told them I’m going into nuclear six months ago, they were all like, or seven months ago, they all said, Trey, are you crazy? Are you all right? Are you having a midlife crisis? And now it’s funny, they’re sending me articles saying, hey, you you might be onto something here. This is an industry that you can learn.

come into it with a lot of humility, come in saying, I know something about tech or I know something about construction. I want to learn nuclear. And you will find there are resources about, there are podcasts, Mark, like your podcasts. There are people that are very welcoming. So think the biggest message is if you are new into your career or even if you’re multiple decades into your career, if you’re looking for a career change, nuclear is the place to be. And it’s going to be a very exciting

space over the next five to ten years if not 20 years because look maybe going full circle there’s very few bipartisan

Laws and bills getting passed the advance act passed 88 to 2 Republicans Democrats Everyone’s agreeing nuclear is the path forward So come into this space because it’s about to be we’re about to see billions and billions of dollars of construction of new projects This is the place to be so my message will be come and join I know it’s kind of scary like coming in but there’s a lot of great people here here to guide and any Entrepreneur out there if you’re new to nuclear

I would love to take a call and help in any way I can. So that’d be the parting words.

Mark Hinaman (56:58)
Awesome, Trey Lauderdale couldn’t end on a happier and more optimistic note than that. That’s fantastic. Thanks so much for your time, yeah.

Trey Lauderdale (57:06)
Of course, Mark, thank you for all you do. Enjoy your time in Romania and around the world and look forward to this podcast and reviewing your future podcasts that come out. So thanks for everything you do.

‍