The Exchanges

Q let's, I think we should start by giving diesel generators their flowers. There's a reason why we use them everywhere all over the world. So we should talk about what's good and uniquely good about them, and then we can talk about what's bad, which in my opinion is a longer list. But first, like, what is good about diesel generators? Why are they fit for purpose for these applications?

A Yeah, there is a very good reason that they are so prevalent and nothing has been able to displace them so far, because it's a very unique combination of capabilities. So, number one, a diesel generator is essentially a highly convenient microgrid in a box. When you have the generator, you need nothing else to provide electricity wherever you are if the grid is down or if the grid's not available. It's also incredibly low cost. Diesel generators are on the order of hardware alone, five to 800 dollars per kilowatt. Even the more expensive emissions controlled ones are about a thousand dollars per kilowatt. You're talking about incredibly low cost compared to a gas turbine or supplicating engine. Or a fuel cell or any other, most other kinds of technologies. And then fundamentally, it's also about the fuel itself. When you think about what diesel is as a liquid fuel, it's a way to store energy incredibly compactly. Ultra high energy density. It's reasonably stable. It won't just, you know, catch on fire on its own spontaneously. It doesn't leak out. You don't need specialized tanks like you do with natural gas or other gaseous fuels like hydrogen. Um, so it's a very simple fuel Easy to store on site. You have a simple plastic tank. You can store 48 hours of fuel or 96 hours of fuel. What people in the battery world can consider very long durations is basically trivial to store wi…

AI assessment note: “number one, a diesel generator is essentially a highly convenient microgrid in a box.”

Q What is, in, in the modern day, but prior to, you know, revolutionization by AI, what is utility resource planning actually look like inside the utility?

A Yeah, there, there's many parts of planning. When you say resource, it might be integrated resource on the bulk power system, or with, um, centralized generation, there's transmission planning , there's distribution planning, then there is DER planning. Each one of these are on silos, um, and recently just talking to a very large East Coast IOU, even transmission planning, there are like 12 departments doing their own thing. And each time a study is done, it's completely separate and siloed, so each use case is like a study. You want to connect a generator, it's a study. If you want to connect a load, it's a different type of study, but the underlying model is still the same. So studies today, I would say half if not more of the time, especially for distribution, is on cleaning up data. So data is in a ton of different places. Data quality is not too par, and a lot of manual effort is required to pull data together. Then you run the analysis, which is really tuned for a worst-case scenario planning. What's the five hours of the year that's going to be worst case for the next 10 years, and you plan to that level of standard. Now, people are improving, like looking at 96 hours per year, two 88 hours per year, five 76 hours per year, so each of those are like high, best case and low cases on weeks of the season or month over the years, and I would say the gold standard today is pr…

AI assessment note: “Each one of these are on silos... half if not more of the time is on cleaning up data.”

Q What can you actually prefab? Like what I'm trying to, what are the Lego building blocks?

A Yeah. Yeah. Good question. So, um, almost everything in the electrical world and almost everything in the mechanical world are prefab for us now. So think of all the way out from where the utility comes onto the campus and I've got a transfer switch or I've got a transformer that I'm stepping down. We'll have all of that pre-assembled on a, on a skid so that you don't have to have somebody wired between the transfer switch and, and Wire between that and the, you know, a transformer. Um, we'll have the entire power center, you know, everything in there that distributes electricity into the building all show up in one unit. Um, we have our entire, uh, today, both liquid or hybrid cooling, so both liquid and air, shows up in a single packaged unit. So these things, you're not having to put together any of it on site. Um, all of our buildings are tilt up You know, um, prefabricated so that the panels show up and we literally just crane them and set them in place so no one's having to make any of the walls, no one's having to pour any, uh, forms or, or build any forms or pour any concrete. They all show up on a truck, again, literally just like Legos. Slab A goes in slot A. Slab B goes in slot B. And, uh, they get lifted off a truck and slid into place. So almost the, uh, the roof is the same way. Uh, we have prefabricated double T's that support the span. Literally the whole buildi…

AI assessment note: “almost everything in the electrical world and almost everything in the mechanical world are prefab”

Q at, you amortize it across a much, much smaller project, and that blows out the economics and it kills you. Um, this is where I think, you know, the way that you've been thinking about it is the most interesting to me, because that's the, this is the problem that's been the most intractable in my mind. So what, what gives you confidence the soft costs are going to change?

A Yeah. All right. So, If you think about sort of the total project cost of a C&I battery, right, roughly half of it is what we traditionally call soft cost, right? So let's just say for the sake of argument, uh, 800 dollars a kilowatt hour is sort of the average C&I project in the world today. So traditionally, the way to think about it, and this is sort of post-ITC, right, is three to 400 dollars of that is the hardware. About a hundred dollars a kilowatt hour of that is the soft software. So 400 dollars a kilowatt hour gets you the hardware and the software obligations you need, right? Um, the other 400 dollars a kilowatt hour is what we traditionally call soft costs, and that's broken down into two buckets. About 200 dollars a kilowatt hour of that is installation costs, and about 200 dollars a kilowatt hour of that is what I generally call transaction costs, right? Which is all the things you mentioned. Interconnection, permitting, financing, You know, the people, cost of customer acquisition, all those types of things. Um, Like, I think there are huge opportunities for reduction, both on the installation side and on the transaction cost side. The thing I've really been focused on over the last, you know, 18 months is the transaction cost side of things, and that's where I think AI is a huge part of the solution, right? And so, you know, the reality is, is that a lot of the …

AI assessment note: “train agents to do those repetitive workflows, and that drops your transaction costs significantly”

Q All right, so we're going to talk V to G. I want to start from the technical perspective. What's actually required for a vehicle to discharge its battery into the grid?

A Yeah, there's three key components. The first is a bi-directional electric vehicle. It has to have the ability to draw power from the battery and send it to buildings or to the grid. The second piece is you have to have a bi-directional charger that can both send power to the vehicle and can also receive power and send it through the charger back to buildings with the grid. And the third piece is you need Uh, some sort of software platform that manages and optimizes that power flow. It responds to constraints that are established by, uh, the customer, the owner of the EV to ensure that the vehicle, uh, is first and foremost providing the mobility services, uh, that, uh, the vehicle was purchased for from the consumer. So those, those three pieces. And, uh, you know, we can take it, uh, another, uh, level, um, particularly around where the conversion Of the power from DC. We know we take DC power from a battery, and we convert it to AC power for use in buildings in the grid, and there's, there's two different flavors of architecture out there, and I'm happy to talk about that if you think that would be of interest.

AI assessment note: “there's three key components. The first is a bi-directional electric vehicle.”

Q Yeah, let's do that. Talk about how you can configure it. And then, and then we'll talk about sort of like where we are today in terms of those. Do we have those three things out in the wild and where?

A Yeah, absolutely. So there's, um, what we call a DC and an AC architecture, and that really defines where the conversion, uh, in a device called an inverter, which is a grid interactive inverter. It takes the DC power and converts to AC power. As I mentioned before, for the DC architecture, that conversion happens Offboard the vehicle. It's either in the charger or a wall box that's adjacent to, um, to the charger. So it's offboard. So it's drawing DC power from the EV, converting it to AC power offboard and sending it to the building or to the grid. And then, um, the other flavor is V to G AC, and this is where their onboard, uh, conversion occurs. So the vehicle sends Uh, AC power from the vehicle to an AC bidirectional charger. And so just to give you a sense of, uh, where we are commercially, the, the Tesla Cybertruck, uh, V to G product is an AC architecture. So there's an onboard, uh, inverter capabilities and the Ford F- one-fifty and the, um, GM, uh, suite of EVs that have V to H and, uh, increasingly V to G are based on that offboard, uh, V to G DC architecture.

AI assessment note: “just to give you a sense of, uh, where we are commercially, the, the Tesla Cybertruck”

Q dynamic in the market is that because there is so much of a supply constraint in providing power or generation or whatever, um, you know, the, the supplier can demand more out of the customer. So yeah, is that becoming a challenge for the NeoClaus? Is it putting them at a competitive disadvantage in being able to build capacity relative to like the hyperscalers who obviously have big balance sheets?

A Yeah, and a pretty massive one. Um, you can just look at the numbers, you know, CoreWeave, they have, you know, three and a half gigawatts of contracted power. Contracted for that means signed leases for the most part, some self-built, mostly signed leases with third parties. And what you saw was that this number was about, if I remember correctly, 1.3 gigawatts in Q four, 20, 24. Uh, so they've scaled that up pretty fast. Uh, but since Q three, 25, they haven't really been able to secure more. And that sort of coincided with the overall tightening of financial conditions, where you saw a pretty massive bond sell-off, which impacted the likes of CoreWeave, Oracle, and many of these guys. And suddenly, sort of, the high-yield market froze to some extent, right? And that's also, you know, obviously related to the fact that this market is not that big, and basically, you know, they massively increased the supply on that market. Anyways, we get, we sort of get to where we are today, which is that it's getting pretty tough for these companies to get the financing Um, for all of these parts. And they're all, as you said, more and more capital intensive. Utilities now are asking multi-billion dollar commitments for gigawatts, uh, gigawatts of power. Turbines and so on and so forth is the same thing. So yes, pretty massive disadvantage. And again, like I go back to what I said earlier,…

AI assessment note: “Yeah, and a pretty massive one. Um, you can just look at the numbers”

Q and, and those who import into the U.S., and There's all the standard versions of differentiation. The product is different. Cars are different from each other. Um, what do you see as, like, the biggest differences amongst the major players, at least in China? Are they approaching the market with different strategies or different types of products, or are they all pretty similar and they're just cutthroat with each other?

A Oh, there, there's, I think of them as, you don't ever see that picture of, like, a massive school of Schools of fish all mixed together, shaped sizes, colors, going in different directions. That's China's auto industry. So you have niche players, you have mass scale players, you have electric only specialists, you have companies that will produce every powertrain. But if I were going to think, try to distill it down, you have really, uh, Old generation or your legacy automakers who are dependent on scale and low cost. That's your BYDs and GLEs and SAICs. Then you have this new generation. That's all about software and software defined vehicles and over there updates and digital interfaces. So they're focused on making the car gadget, the urban, the ultimate urban device. And so you have. Kind of, if we put it in American terms, you have Ford, GM, and Chrysler on the one hand, and you have Tesla, Rivian, Lucid on the other. That's the biggest distinction, uh, relevant to us in our discussion today.

AI assessment note: “Old generation or your legacy automakers who are dependent on scale and low cost.”

Q to the U.S. then, in the rest of the world, you know, these Chinese cars are flooding the market. Are we seeing automakers outside of the U.S. start to feel the pain meaningfully? Like, are we going to see a wave of failures of automakers of basically everybody who's not Chinese and not U.S.-based, or at least not importing enough into the U.S. for it to make up the difference?

A We're already seeing it in Europe for sure. Uh, Volkswagen's come out and said between now and 2030, they'll lay off 50,000 people. This is unprecedented. They're closing plants for the first time since World War II. And the Germans are not alone. Honda, also for the first time since the 19 fifties, uh, announced a loss last year. Why? They're getting beaten down in traditional markets like Thailand. Indonesia, Australia, the UK, across Europe, uh, virtually every legacy automaker is taking it on the chin, and the most vulnerable right now, there's two of them. One is Nissan, and the other is Stellantis, the group that owns, you know, Fiat, and Maserati, and Jeep, and Chrysler brands. Those two are the most vulnerable, so what will happen is either they'll joint venture with the Chinese And eventually be taken over by them, or the Chinese would just buy their brands outright. I mean, we're not talking about something that might happen three to five years from now. It's happening right, right now in Europe. Enormous pressure on the existing automakers just to stay alive.

AI assessment note: “We're already seeing it in Europe for sure. Uh, Volkswagen's come out”

Q that you mentioned, they're selling at 10,000 dollars. Even with a hundred percent import tariff, they're still going to be cheaper in the US. Like, if you're trying to buy an EV in the US, it's hard to find a 20,000 dollar EV as well. So I'm sort of surprised that even despite that tariff, we haven't started to see a lot of imports. Why do you think that is?

A I really appreciate it at the outset when you said we might go deeper in the weeds because it gets a little bit hairy here. So a couple of caveats that, that under 10,000 dollar vehicle that's sold in China would not qualify for registration United States on safety. And probably on overall emissions, depending on if it's PHEV or EV, but mainly safety, homologation is not there. So, for example, that same seagull that sold in China at 10,000 dollars, let's call it, sells in Mexico for 20,000 dollars. Mexico's standards are not as high as the United States, so we're talking about 25, 30,000 dollars. Now you have a small car with small range, and guess what? The American consumer goes not really into that. I know it's 25, but it's small and short range, ah, not that great looking. Uh, no, I want a bigger vehicle. So let's be care, I think we need to be careful about what kinds of cars would actually, the Chinese would actually aim to sell here. Second thing to note, the Chinese are starved for profits, and when I talk to the Chinese manufacturers, they're far less interested in solving America's affordability program And far more interested in saying, how can we attack the larger SUV and pickup truck segments? BYD, for example, recently launched this midsize pickup truck called the Shark. It's aimed directly at the Tacoma and the Ranger, and I know other Chinese automakers have th…

AI assessment note: “under 10,000 dollar vehicle that's sold in China would not qualify for registration”

Q autonomy for a second. Where, where does autonomy fit into the mix in China? Um, you know, obviously in the U S we have, we have Waymo, we have the Tesla robotaxi coming. There are others like Zoox who are starting to arrive in some cities. Is China seem ahead behind? Is it, uh, is it part of the export strategy or is it not? Like where does autonomy fit?

A Great, great, uh, arena here. Autonomy. So it is a two horse race. Between the United States and China, the edge to China comes on the regulatory front. Here's a comparison. When I talk to regulators in China, their job, as they see it, their mission is to smooth the way to commercial ramp of autonomous in China and other markets worldwide. Make it as easy as possible. Facilitate. Make it go faster. What do you need to make it happen? Whereas here in the United States, the regulators are much more cautious. How do we make sure things don't go wrong? How do we make sure it's safe? So in that respect, The U.S. leads in technological innovation, but China, once again, is quicker when it comes to commercialization. Not only inside China, but we're seeing the Chinese autonomous vehicle makers move into the Middle East, and in the U.K. and Germany now, moving quicker to market than their U.S. counterparts, and this is where we're at risk. We might be the inventors, but China commercializes more quickly, and that's a problem for us.

AI assessment note: “China, once again, is quicker when it comes to commercialization.”

Q Okay, let's get a little closer to home, um, and start with the, our neighbors to the north and to the south. So you mentioned Canada got rid of its hundred percent tariffs. You mentioned Mexico still has a 50% tariff. So in both of those cases, how quickly and how much have Chinese vehicles started to penetrate those markets?

A Okay, so let's take Mexico first, because there's a little bit of history there. Mid-COVID, the government decided, oh, Our consumers need affordable vehicles, so we'll drop import duties to zero, and that was the beginning of this torrent of Chinese cars going into Mexico, and for the last two years, Mexico's has been the single largest destination of Chinese exports. So much so, the United States said, hey, what the hell's is going on down there, guys? You're letting all these Chinese cars, and I was recently in Mexico City, you see dozens of brands with, there's Chinese cars all over the place, dealerships going up. And so the U.S. put pressure on Mexico finally in January of this year to raise that import tariff from basically nothing to 20% last September to now 50%. So we've seen a deluge or an outpouring of cars into Mexico. There's several 100,000 on the road now. Many brands. It's slowed a little bit by the tariffs, but guess what? The Chinese are now Eyeing plants inside Mexico. They'll do assembly and sell to Mexican customers. So Mexico is a strategic launch pad for the, for the Chinese in North America. And if we go north of the border, just as you said, Canada held the line on imports at a hundred percent with the United States until earlier. Gosh, time flies so fast. Earlier this year, when Carney went and said, we're going to allow our first 49,000 cars from Chi…

AI assessment note: “There's several 100,000 on the road now. Many brands.”

Q understanding is that there are various different versions of that. They did it at the, at the reactor level, but not, uh, not a plug break-even, or whatever other term you want to use. So like, or orient me, what, what did that Actually show, and then what's different about that from what we're going to have to show when we want to turn fusion power plants into something real?

A So what they did, and you're gonna have to forgive me if I, if I get some of the details wrong, I'm not a physicist, I'm an operator in the field, so disclaimer. Um, what they did was they stored about 300 megajoules of energy in a capacitor bank. They used a laser to drive about two of those megajoules into a target, and then they got five megajoules of energy out of the target. So really, really big achievement. They've since improved on an achievement, gone closer to eight megajoules out of the target. And so if you draw your, your proverbial box around that fusion target, then you got more energy out of the target than you drove into the target. But to your point, if you draw your box around the entire fusion machine, including the capacitor bank, you only got, you know, percent and a half or so of the energy stored in the system out of the, out of the machine. And obviously that's not a practical basis for a power plant. You need to get about five X more out of the machine than was stored in the system to have a practical basis for a power plant. And so, to your question about milestones, that is the next big milestone for this field. It's demonstrating something called net facility gain, which is getting more energy out of the entire machine than everything required to run the machine. We like to define it as all of the energy stored in the system. Um, and a number of com…

AI assessment note: “demonstrating something called net facility gain, which is getting more energy out of the entire machine”

Q you mean? Obviously, if we're talking about fission world, there's like a, such a broad spectrum. If people talk about modular, they could be talking about a microreactor that's a megawatt, or they could be talking about, like, versions of SMRs they call modular, uh, and that's literally in the name SMR, but it's still 300 megawatts plus. So, like, what's a, what's a pixel size for modularity for you?

A I just mean big things that themselves consist of many small things. So tabletop scale fusion doesn't work, but our goal is to build fusion power plants in the couple hundred megawatt range, so in the two to 300 megawatt range, that themselves consist of modular mass-manufacturable building blocks. Um, and so in our case, the, most of the capital cost and footprint sits in the driver. This is true for a lot of different fusion approaches. For us, that driver consists of a 156 identical modules. Each of those modules produces more than a terawatt of peak power and sits in about the footprint of a shipping container and is made from oil, plastic, metal, and water. So we bring two things that make a big difference, right? The first is an established scientific foundation based on decades of work at the national laboratories and the breakthroughs that I mentioned at the beginning of this conversation, but also a path to a modular, maintainable, deployable system that can scale more readily as a power source.

AI assessment note: “156 identical modules. Each of those modules produces more than a terawatt”

Q Is it, is it in that case because they have insufficient supply of gas to feed those, those projects, or is it that gas prices are too high and so they'd be losing money on producing more?

A It's gas prices too high. Uh, when you look at the economics of it, it just doesn't make sense with where global nitrogen values are. And I also think there's the political aspect of it, right? The European political engine has been really charging towards green energy and Frankly, this old school, old technology nitrogen production, from their viewpoint, is very dirty. It's outdated. We don't want to have anything to do with it. So you've already got bad economics for the plan, and then you've got a political outlook that basically sits there and says, don't put any money in these plans because you're probably not long for this world. We're not going to support it. We're doing everything we can to shut you down.

AI assessment note: “It's gas prices too high. Uh, when you look at the economics of it”

Q Alright, let's start by having you give me a walkthrough of the, uh, uranium fuel supply chain, the nuclear fuel supply chain. So, like, take me from soup to nuts. What do we start with, and what do we end with?

A Yeah, happy to. Um, I mean, the, the background is that every, every reactor needs fuel, as, as most people know. And we can talk about types of fuel, but, but all fuel in reactors in the U.S. today is made using a five-step process. So, step one is you mine uranium out of the ground. You then convert it to a gas. That's called the conversion step. Um, you then enrich it, which is really a refining separation step. You then deconvert it into a solid, back into a solid. And with that solid, you then make fuel, fuel fabrication. So fuel pellets or trisoparticles or whatever that is. So five steps total. Um, the US does all of the steps. The US does not do the middle step at commercial scale. So that's where the bottleneck is, which I'm sure we'll talk about today.

AI assessment note: “all fuel in reactors in the U.S. today is made using a five-step process”

Q over time as the U.S.-Russian relations have moved? Like, I get the sense it's one of these areas that, like, We kind of don't like to talk about it because we're sort of reliant on Russia to some extent right now, but we need it, you know, and so we're sort of unwilling to sanction it or stop buying from Russia. Is that, do I have that sort of right?

A Well, in, in twenty-twenty-four, there was a Russian uranium imports ban passed by Congress, and so there's a waiver process, um, that's, that's ongoing right now where the Secretary of Energy can waive, uh, the ban. If a utility needs it and there's not another source, which has been the case, um, that waiver process expires January first, 20, 28. And so the setup today is yes, it's still three quarters Europe, one quarter Russia. Um, most of that Russian uranium is coming in. It's all coming in under those waivers. Um, I think it's gone from about 25% to 20% as utilities look to diversify and get ahead of the full 20, 28 ban. But That is currently the breakdown. Um, a lot of people have asked, how, how do we even get here? How is it the case that we're still importing from Russia? You have to go all the way back to the fall of the Berlin wall, the end of the cold war. So eighties, the U S was the leader in global enrichment, something like 86% at the peak. Um, and then the Berlin wall fell and we entered a treaty with, with Russia, which was called Um, the megatons to megawatts program, and in that, in that trade program, we imported Russian, uh, warheads. We downblended them and used, used that downblended material to run our reactors. Um, we then, you know, sent the depleted uranium back to Russia, uh, to be, or the, we sent the depleted uranium back to Russia to be enriche…

AI assessment note: “in twenty-twenty-four, there was a Russian uranium imports ban passed by Congress”

Q Okay, so significant differences, as you said. Maybe can you help me break it down as I think about, um, the, all the different reactor types that are currently being pursued to, to come? Like, which ones present the easiest waste handling challenge? Which ones present the hardest?

A Right. So anything that is a high-temperature gas reactor that nominally uses Triso, and there's a lot of these that are in the mix there, right, that would not be a challenge. Any of your molten salt reactor designs that use a Triso fuel, that would not be a challenge. If you have a molten salt reactor with fuel dissolved in core, that would be potentially something that you need to address and condition and things of that nature. Any of the sodium fast reactor designs, so your TerraPower, your Oklo, things of this nature, this would be A conversation where you need to have, at a minimum, a waste conditioning component to this, and if you're a company like Oklo, that's very much thinking openly about recycling, right, you're, you're basically moving down that path anyways, so, um, and for any of the light water reactor designs that are being considered, like, uh, the GE Hitachi reactor, um, Westinghouse, right, that's not an advanced reactor design, and when the AP 1000, right, we have very established pathways with respect to managing those types of

AI assessment note: “anything that is a high-temperature gas reactor that nominally uses Triso... would not be a challenge”

Q go back to the grid, then, just to wrap up. Over time, and obviously this will take a long time, but if over time, if we go and start to, one by one, go throughout the transmission distribution system and replace all of these traditional oil-filled transformers that are on the grid right now, ultimately with solid-state transformers, like, big picture, what does that enable from a grid management perspective?

A Well, utilities and grid operators right now are facing, um, a lot of pressure, right? They've got aging infrastructure, growing demand, um, and they, they're in the market for new solutions, and luckily, SSTs can provide a ton of value propositions beyond just voltage transformation. Um, an SST can have a cost similar to a traditional oil-filled transformer, um, but at the same time provide functions that would be That would be provided by popcorn components around the transformer. Functions like overcurrent protection, fault isolation, what an automatic tap changer does for voltage correction, uh, what three phase balancers do to enable higher utilization on the different phases in the distribution grid. They can provide the spinning inertia type functionality that synchronous condensers do for frequency regulation. Um, and they can also take the place of cap banks for power factor correction. So With the choice to go SST the next time they need to place a, uh, a distribution substation down or replace an aging, fifty-year-old, you know, 34 KV to two away transformer. They're at the same time getting all of those other value added functions kind of for free. And what those other value added functions do is enable more utilization of the existing poles and wires. And utilization is the key to affordability. Um, if you look at the rate cases, uh, uh, for public utilities, uh, a…

AI assessment note: “SSTs can provide a ton of value propositions beyond just voltage transformation.”

Q in that order. So let's start in the mid-Atlantic. I want to get to this DOE emergency order and the, you know, concept of data centers bringing their own generation and all that, but I want to get there after we talk about what's happening in that market at the high level. So start me off with the 30,000 foot view. Like, what is the state of affairs in PJM?

A Yeah. Look, there's been a lot of change, and the change has been quite rapid. When we rewind back just a couple of years, we had capacity clears that were going off between 30 and 50 dollars per megawatt day, and that compares to a little bit over 300 dollars a megawatt day in the more recent clears, uh, and subject to the cap in the capacity auction mechanism. Those low clears were in many ways, I think, sending the signal that we had more than enough capacity to We had a period of time between, let's say, 2008, the great financial crisis, and, ah, really the last year or two where there was virtually no demand growth in the PJM market. And in the absence of that demand growth, the environment was really one where new assets were being added to the market to replace older assets, less efficient assets, higher fixed cost assets. Uh, along with the addition of renewable projects supported oftentimes by state-level mandates or corporate procurement. So in that environment, there wasn't a need, and there was certainly no price signal to go actively develop and build new generation outside of renewables. And that's the backdrop really Kind of shifted really quickly over the last couple of years as probably a combination of factors hit the market and has driven up demand growth from that very quiet, virtually no demand growth, almost a market stasis to the real need for new generat…

AI assessment note: “capacity clears that were going off between 30 and 50 dollars per megawatt day”

Q It seems like a lot of the focus, both from a policy perspective and also a market perspective, is on, let's build as much new gas generation as quickly as we can to meet the capacity need in BJM. And it sounds like you're saying, to some extent, yes, but that actually isn't necessarily the, the savior, especially in the near term. Is that right?

A Yeah. When I look at the resources that we have available, Large scale gas is a 20, 30 plus new resource. In the interim, there are other things that we can do. We can add demand response. We can add batteries. We can upgrade existing facilities. We're working on the conversion of some of our combustion turbines to combined cycle power plants that can be done more quickly than building something that's completely de novo. We have a project where we can swap out combustion turbine blades and get dramatically more capacity. So there are different things that we can do as a bridge to large scale, completely de novo, new generation. But if that's what we're hanging our hat on, we need to anticipate that that's an end of the decade plus project. Deliverable at large scale resource.

AI assessment note: “Large scale gas is a 20, 30 plus new resource. In the interim”

Q our company. And that is, you know, Indicative of progress. But, but I have no idea whether that's indicative of progress, right? Like, does that happen every single time? Is it always positive? Do people get rejected somewhere in this process? Like, what is it like in that murky, that murky early ground where there's all this iteration and, and work being done before you submit your formal construction permit?

A Yeah, absolutely. Um, I would say that while it is an indicator that things are happening in the company and an indicator that the NRC is not in the dark about what's happening in that company, at no point, sort of, in that process is NRC really issuing official decisions of any kind. In the pre-application activities, you do have to submit kind of a letter of intent. You have to submit, you know, billing information so that the NRC can start billing you for their time. Uh, and you, You do things like a regulatory engagement plan gets put together, and you tell the regulator how you want to, like, go about the process of engaging with them, and then there's usually public kickoff meetings and readiness assessments as a part of that, including public outreach meetings, and all of that is before you submit your application. Once you submit your application, NRC then decides whether or not to accept That application and start actually doing the review. People do get rejected at the acceptance of application stage, for example, if it just doesn't reflect the level of technical depth that NRC is going to need to do the review. That is certainly a thing that happens, and sometimes it's, you know, years after you've started your pre-application activities.

AI assessment note: “People do get rejected at the acceptance of application stage, for example”

Q the state of affairs? If I want to, if I want to plug one of these little batteries in, in my house, uh, maybe I'm using it as backup for my fridge or whatever it might be. Um, what is the lay of the land in terms of what requirements it needs to meet? What permissions I do need? Is it murky? Is it defined? Like, where are we there?

A It is murky. I think what really matters is the thing we focus on a lot is safety. So there are plenty of UL certified products that, that adhere to the NEC, uh, saying you can plug this in in the following manner and it's safe, uh, to do so. And so there may be jurisdictional like AHAs or DOBs or fire departments that have an opinion on what should go in a given location, how big of a battery or something like that. But at the sort of electrical code level, these are, this is already allowed under the current guidance, and there are many products that support that. So from that lens, you could say in most places you can go out and buy these things and, and plug them in in whatever state you're in. Um, a lot of the attention that's happened recently around regulations is specifically there's bills now introduced in I think it's up to 30 states, or sorry, 24 states with maybe 30 soon, um, Introducing bills where you can actually export to the grid through these devices. And so we think of that as an extremely important distinction where a lot of that regulation that's being passed is focused on really what is an interconnection agreement? What permission do I need from there for the utility? Whereas I understand the utility's concern is, hey, if you just start exporting and the grid goes down and our line workers out there, they don't actually know a line is, is energized and, a…

AI assessment note: “It is murky. I think what really matters is the thing we focus on”

Q like, really try to understand. The topic basically being how much of inference compute Might move from central cloud infrastructure to the edge, and then how far to the edge, of course, being another question. I think we should start by actually defining those categories a little bit. How do you think about the categorization of, like, where compute can occur? Then we'll talk about each of those categories individually.

A Right. So even before we talk about generative AI there, for classical compute, cloud computing in general, All of the services we loved and changed the way we live and work today. There are three levels generally I think about for compute. The first is massive hyperscale data centers, the ones run by Microsoft and Google and Amazon, hundreds of thousands of machines, massive facilities. That's what most people think about when they think about cloud computing. At the other end of the extreme, Would be a personal devices, consumer electronics. So you think about your phone, you think about your tablet, uh, your, your, your laptop, uh, plenty of compute can happen there as well. There is a perhaps less understood, uh, middle layer or intermediate layer called edge computing. And edge computing really means that there are times where you don't want to go all the way to this remote massive facility, uh, And wait for the data to go out to that data center and then come back. You might want to access some compute that's a little bit closer to you, maybe in the same city, maybe in the same geographic region, that's edge computing. So they're still going to supply really capable, high-performance machines, these servers, um, but you don't suffer those longer communication times or latencies that you might if you, um, if you were to go to that remote, massive data center.

AI assessment note: “There are three levels generally I think about for compute.”

Q here. But then it Seems to me that because AVs were generally delayed or maybe the need wasn't as high, like what we've got today, if you just look at the infrastructure today, it's seems like the vast, vast majority of classical compute even, um, except for stuff that's sitting in like mainframes at companies is in the cloud and the big centralized data centers. Do I have that right?

A That's right, and this is a decades long trend. I mean, we've seen this progression, uh, this adoption of cloud computing over the last 15 to 20 years, and there are a couple of reasons, uh, we are seeing that shift, or we have seen that shift. Uh, the, the first is that, uh, Computing in a massive data center run by the hyperscaler companies, these big tech companies, is much more energy efficient. They know how to deploy these facilities. They know how to cool them and build HVAC systems, um, efficiently. So they're incurring very, uh, very small overheads per watt of compute. There's this industry standard metric called, uh, power usage effectiveness or PUE. And that's the ratio of how, of the power you're using divided, compared to the power that's going to compute. So Google's PUE is close to 1.1, which is to say for every watt going to compute, there's an additional .1 watts going to the overheads of power delivery or cooling or whatever. So that's really incredibly efficient. And most mom and pop data center operators, most enterprise data center operators don't get the scale and efficiency that these hyperscalers do. Um, the scale also gives a second key advantage, which is the ability to share hardware. So you buy the hardware once, and you have lots of users sharing the same physical hardware. That allows us, again, to drive the cost down, allows the hyperscaler opera…

AI assessment note: “That's right, and this is a decades long trend.”

Q that is the next wave of applications for AI, right? And so maybe we go back to the autonomous vehicle world and things like that, where like latency making decisions in near real time does become really important. Robotics being another category that could be a major user of AI compute, but needs really, really low latency. Is that part of the argument for shifting some compute to the edge?

A Yes, absolutely. So The class of compute you mentioned autonomous vehicles, robotics fit into what we call cyber physical AI. So cyber physical systems are those that have a cyber component, a computational component, but also interact with the physical world. And once those interactions with the physical world arise, then we care about responsiveness. Because that underpins safety guarantees, and the ability to make sure that your robotic arm is able to respond quickly enough to hazards, your autonomous vehicles are able to do so. So I, I agree that there will be cases where we will need those really low latencies, and that is going to require edge computing much closer to the user, so we have much shorter internet delays, network delays.

AI assessment note: “Yes, absolutely. So The class of compute you mentioned autonomous vehicles, robotics”

Q hard to do, and indeed is, but these are all hard problems. So if that happens, Do you think that we are going to see a significant portion of that inference workload move to that type of scale? Is that the right scale? Like, should we be looking at 10 megawatt sites, a hundred megawatt sites, one megawatt sites? Like, how far to the edge do we want to go?

A Yeah, absolutely, and I, I agree with the premise of that question, 100%. I think that there are two reasons to go to smaller, many smaller data centers. The first is the one you mentioned, power, power provisioning, uh, and connections to the grid. The second is, uh, the fact that you don't need massive GPU coordination for an inference workload. Um, I, I guess the catch might be that if you are thinking about Your existing edge data centers. Maybe you've got data centers in downtown Los Angeles or something like that already serving workloads. Those workloads may not be configured to handle GPU and AI compute. Uh, they may have, uh, power delivery infrastructure that was optimized for CPUs. They might have, um, HVAC systems optimized for the much lower power density of CPUs. So it's not simply a matter of Pulling out your CPUs and replacing them with GPUs. You're going to, you may have to retrofit the, the facility itself to support that. Uh, but I, I agree. I think finding capacity there may eventually become easier than finding the next, uh, thousand megawatts.

AI assessment note: “I think that there are two reasons to go to smaller, many smaller data centers.”

Q Yeah, and as a result, you've been pretty diversified in your approach, focused on everything from fusion to geothermal to carbon capture to hydrogen. How would you characterize this overall approach? Portfolio strategy.

A I think we're really trying to take a practical approach at decarbonization and understanding that you have to have energy intensity as well for certain types of manufacturing, other types of activity, but that also means we've passed the strictest oil and gas regulatory regime really in the country, if not the world, here in New Mexico. So we've got the lowest carbon intense barrel of oil and unit of natural gas anywhere. That was where we started, because that's what the energy mix we had already. We're number two in oil production, number three in natural gas production in the United States. We have the only uranium enrichment plant in the United States today is in New Mexico. We have the most productive geothermal well now operating in the US happening in New Mexico. So we're looking at what are those needs around firm load in particular? Um, we've done great on, on wind, we do great on solar, but also where can we invest R&D side on, on deployment side? That's going to create that firm load that really creates the opportunity for reshoring for, uh, whether it's data centers or other heavy users that allow us to grow our economy while ensuring we're meeting our climate goals as well.

AI assessment note: “we're really trying to take a practical approach at decarbonization”

Q And if I understand it correctly, New Mexico has the nation's second largest sovereign wealth fund. How big is that fund, and how, how is it directly investing in, like, deep tech and energy innovation?

A Steven, it, it, it's very exciting to talk about what the SIC, the State Investment Council, is doing. So, uh, it, it, it's now about sixty-five billion dollars in this fund. It's, it is the second largest in, in the U.S., only behind Alaska, and we think we'll probably surpass Alaska in the next three years or so. What they've done is they've taken about two billion of that and put it into a venture fund that is working with the 25 to 30 largest deep tech venture funds in the world. Think groups like Lower Carbon, Antler, Colesla, Up, and they're investing in those areas that align with the state's strategy. Our strategy, where we think we have the right to win in New Mexico, is around advanced energy Advanced computing, edge computing, space and defense. These are all areas because of Sandia National Lab, Los Alamos National Lab, Air Force Research Lab, all located here, our universities. We have a unique competitive advantage. So now the State Investment Council is funding these funds who have portfolio companies who are aligned with that strategy. Think Pacific Fusion. Pacific Fusion It has three different funds that are funded by the State Investment Council funding that company, and part of the direction of the State Investment Council is give us a return, but also deploy capital or create jobs in New Mexico. There are very few places in the world that have created this d…

AI assessment note: “it's now about sixty-five billion dollars in this fund.”

Q So you've got this capitalization strategy. How does that pair with tax credits, R&D tax credits, jobs credits, equipment credits? Like, how do you put this package together?

A That's a great question. And if you think about What happens with the State Investment Council, right? They're taking state dollars, putting them into company as equity investors. So now my job is to help that company be successful in New Mexico, because that's going to give us a return to our investment, but it's also going to create jobs. It's going to create economic opportunity on the ground. So we, we have a research and development tax credit, which can pay up to 10% on qualified research expenses. We have a high-wage jobs tax credit that's refundable at eight and a half percent of your employees' wages for four years. We're going to pay between 50 and 90% of your employees' wages for the first six months that are in New Mexico on new hires, depending on where you do it and the type of job. And then we have industrial revenue bonds, which can waive property and equipment taxes for up to 30 years. We also will help build capital infrastructure with our Local Economic Development Act Uh, program which could help you build a building, rent a building, or make improvements. So when you layer all of those things on with a capital stack, that means if you leave California, you don't leave Silicon Valley capital, because we're going to make sure those folks are paying attention to you. It creates a really robust incentive structure.

AI assessment note: “when you layer all of those things on with a capital stack”