Try Free

Ep 143: Demystifying Data Centers with Dan Crosby: Energy, AI, and the Future Workforce

Data Center Revolution Podcast June 21, 2026 1h 30m 18,753 words 1 views
▶ Watch original video

About this transcript: This is a full AI-generated transcript of Ep 143: Demystifying Data Centers with Dan Crosby: Energy, AI, and the Future Workforce from Data Center Revolution Podcast, published June 21, 2026. The transcript contains 18,753 words with timestamps and was generated using Whisper AI.

"all right well welcome back to another uh issue of data center revolution and uh dan please introduce yourself for the guests uh dan crosby founder and ceo of legend energy advisors what does legend energy advisors do a whole lot of things so we're across both energy major energy commodities power..."

[00:00:00] Kirk: all right well welcome back to another uh issue of data center revolution and uh dan please [00:00:12] Dan Crosby: introduce yourself for the guests uh dan crosby founder and ceo of legend energy advisors what does legend energy advisors do a whole lot of things so we're across both energy major energy commodities power and gas help customers from you know the inception of figuring out how to get uh energy to a facility all the infrastructure and then whether they're in on-site generation or not then all the way through market risk management it's kind of like cradle two there is no grave on anything energy related for industrial which obviously data center is now fully in the [00:00:43] Kirk: industrial space so overall zooming out of all that how long have you been associated to energy whereas when people are listening and i'm introducing you as one of the greatest authorities in our industry right now in this you've done keynotes at dcac for the last two years um talking about just energy you go an inch wide and a mile deep and you could do that all across the country it seems like right now so your experience stems from years of experience yeah i've been in the space since about 2008 [00:01:09] Dan Crosby: so obviously been major evolutions across markets and commodities you know since then well very good [00:01:16] Kirk: well listen thank you for introducing yourself thank you for getting on a flight i know you're in new york this week but i'm like we have to talk this week about this stuff uh for anybody listening my name is kirk the co-founder and host uh of data center revolution and today's objective right this podcast you've done one with me before some are just to introduce people to the industry um trying to amplify and broaden the aperture to those that have visibility into who we are and what we do as an industry um sometimes we entertain sometimes we educate sometimes we inspire this one is not just an educational one so every one of you that are listening right now in this industry i ask you and i challenge you to lean in and i don't normally do that but i'm going to do it because it takes the chorus to sing louder than the soloist here and what's going to be necessary is if you haven't been paying attention there's a a lot of news about data centers do you agree oh yeah and most of it is not very favorable right and there are three misconceptions that i think that i want to address using this platform and i'm being armed to the teeth by industry luminaries who i'm reaching out to who are sending me just incredible nuggets and bombs about this space and what i see is uh what we are today the the ai revolution or the data center revolution itself is mirroring a lot of similarities between the automobile revolution that took place in the early 1900s and i'm going to pan back and forth but this podcast is designed for you to help me unpackage a lot of very complex and complicated things regarding energy and those that are listening the reason why you want to listen to this and understand it is because it'll take all of us to battle the misconceptions of this industry all of us are dependent on each other to be capable of doing that because we have to protect this industry i'm going to talk about three previous transformational infrastructure revolutions that have taken place and what the pattern was for all of those leading up to the ai boom okay so in the very first one it was the u.s railways and there were two primary concerns when we introduced the railway line the one and the first one was energy where we're getting at the energy to run these trains the second one is what is the impact on our environment right so the second one was when we electrified the u.s grid the number one concern that everybody had in society was how are we going to get this energy to use these lines to transmit energy to other places and what would be the impact on the environment okay fast forward you know early 1900s we really started rolling into this thing called the automobile in between 1919 to 1929 we created 3.7 million jobs 80 of those jobs were created by small businesses so there's a pattern every time that we release something because we released an automobile and and with that we needed infrastructure we needed roads and we needed service stations and in all three of those episodes of transformational infrastructure we needed the government to govern and help us guide policy that made things safer and made things more efficient right think about where we've come from an automobile perspective when we built the car it took decades before we even adopted seat belts and we've come a long way between the time of the vehicle and the time of the ai factory and the concerns about the the vehicles of the times were that they're noisy and they polluted things it was about the impact the negative impact on the communities that they lived in with now you have horses parked right next to cars on the main street and and people were worried about safety general overall safety well the interaction of those two was [00:05:04] Dan Crosby: difficult while the transition was going on too right that's right horses and cars on the same road [00:05:08] Kirk: because the roads weren't adequate enough for both right so same problems we see with infrastructure today same concerns so today the misconceptions regarding this industry are regarding energy water and environmental impact and i want to i think another concern is labor right so sure water energy environmental impact and labor so i want to address those this one i want to this podcast is going to be focused on energy and and i want to start by there's a report that i shared with you that i got from the president of rosen electric actually and it says 226 gigawatts of nothing the physics frauds behind the ai infrastructure boom so let's let's pause real quick before you answer and remind everybody like those are the three misconceptions energy and water primarily being the top two those are all valid concerns do you agree yeah absolutely just like they were valid concerns when we put out railway systems we electrified the grid and when we started making automobiles and roads they were the same concerns then and they're valid to be the same concerns now the same things that we used to get out of that was small small businesses eighty percent of that workforce for small businesses currently today they're without small businesses there is no ai boom ai business uh small businesses represent 75 the entire labor force in the united states right now for building data centers so without us there are no data centers being built so we share those commonalities with transformational infrastructure from behind us and there's mirrors to those right now on an energy perspective the biggest thing that people are screaming about right now is how much base load as in the state of texas as an example it could be up at 8.5 megawatts but we have 220 gigawatt 226 gigawatts in the queue you were talking about this at dc ac on stage in 2024. so help us unpackage this because i'm not saying the author of this one is chicken little but he intentionally deliberately put language in here that was designed to be disruptive or provocative in some capacity and joshua rhodes at the university of texas at austin that was quoted in this we invite you to come be on the podcast not to challenge you to educate us because maybe you're seeing access to things that that dan and i aren't here in the we're in we're in the front lines so we're kind of seeing things i'm not sure what i'm not trivializing in an academic because they don't see what we see i want to see what he sees and i want to know what they know right so you've got a chance to kind of read this and [00:07:42] Dan Crosby: process this help us understand i think yeah i mean there's a whole bunch of things going on right now right you know everybody and their brother is talking about data centers um and you know this has kind of been a trend here for about a year uh i think one of the greatest ironies of what's happening is chat gpt has probably enabled so much content to be put out there um just because it's so much easier to research something than it used to be especially a very complex and nuanced topic like energy and energy markets and so you know the ability to put out things that have provocative headlines that get people's attention and it's not just in articles about you know and obviously fear always sells but even with all these project announcements and all this stuff that's going on you know people are just trying to put out as much content as they can about the most relative topic that's across everything right now and that's data centers um and and really that's data centers because of ai which you know as anybody that started using chat gpt sees that this is you know really revolutionizing everything and i think too one of the interesting things is if you think about the the framework that we have to look at this you know everybody or most people that are in this space and near our age remember what it was like when the internet came along you know and forever it was this huge technology but we didn't know how it was going to be used in fact at our annual meeting two years ago i put up a clip it was a newspaper article from 2002 still talking about how the internet was a fad right and going away and so at that time like we really didn't understand how it was going to permeate our lives and you watch that happen over a period of time and so now looking at this like it's very clear that this is going to do that and far more but how quickly is that adopted you know what is it going to change and there's all these predictives predictions going on right now i think one of my favorite predictions from the dot kind i remember my dad got a computer for the first time i'm a kid and he just went away from a typewriter and he was like deleting things and rewriting and you couldn't see that someone scratched it out anymore and everybody said paper was going to go away right because you don't need to type it up a bunch of times well paper usage went through the roof right like totally wrong so there's so many predictions right now that you know one of my biggest concerns is just people making empirical statements about the future when the future is very much in flux of what this is going to do and every technology ever has done incredibly good things and incredibly bad things take dynamite you know so articles like this is a jillion amount there there's so much stuff on linkedin right now and i think one of the biggest problems is people are trying to make energy seem like this very linear you know linear easy to understand topic and quite frankly it's one of the most complex and volatile markets on earth and so understanding the complexity of that is very difficult and so making empirical statements is an issue and also causation and correlation are not the same thing i've heard this statistic on a audio book i was listening to this morning that over a 10-year period the rise in margin rate across the united states correlated with the rise in divorce rate in the state of maine at a 99 correlation almost one to one wow but most people probably say that the rise of margarine was not causing divorce in maine right so just because something is correlated doesn't mean that it's a cause and that's one of the things that's going on and there's tremendous dynamics going on in energy markets in both power and gas and for those of us who have been deep in energy infrastructure and energy markets for more than a decade there's been a ton of changes over that last decade the changes over the last two years have been wild and a lot of those changes are being made to protect the consumer and they're good things and so there's a lot of things that are having to transition and change right now and and i think for the better there's a lot of uncharted territory that's being navigated right now and a lot of us being navigated very well but you really have to be deep in the weeds and understand how it works if you're you know really going to understand where things are going and how to navigate it well and that's what a lot of people that are you know kind of periphery trying to dive into the data center space right now are figuring out the hard way is this is way more complicated than turning on [00:11:55] Kirk: a light switch i think that all makes sense i think to put it into perspective some people think our industry started in 1994 because that's when amazon.com went live and at the time they represented the majority right you know traffic of the internet and there are less than 10 companies today that represent probably 75 percent of the internet traffic um you know i i would argue that our industry started in 1953 when a bunch of kids the university of pennsylvania and engineering lab working with a software company based in new york created their very first data center it was built for and funded by the u.s government for national defense and intelligence and these kids called that box a mainframe and that would be the genesis of ai in our terms right so our industry's been around for a long time and to date based on what i'm able to discover we have about 55 gigawatts of data center capacity on the u.s grid we we we operate at less than five percent of the grid demand of energy today we are on track to triple in sizes as an industry that's good for those that are listening because we have gotten really efficient at building data centers now the designs a a modern day design data center is the most efficient building on earth and we've rung out a lot of inefficiencies along the way one of them being the demand of water right so we've not only removed the need for water but we've also made it to where if we need water we don't have to use water from a potable system we could use gray water or maybe even brown water in some cases yeah so we're finding ways just like in the automobile the first one through the wall was very bloody and it wasn't very efficient and it wasn't very safe and the demand of energy when we rolled out the automobile the demand for petroleum grew by 50-fold right so we're seeing very similar trends to this type of application of technology emerging and we're seeing a massive demand of energy i've said for years our industry was able to operate in the shadows i mean even in 1994 when we were using data centers more than you were yeah not even until cloud came along then people started using a term called data centers but it wasn't very prevalent and the reason why we are coming into the mainstream 2026 will be the year of the birth of the industry into the mainstream because the automobile in 2026 i mean we're 2026 now 100 years ago 1926 we were building model t's till 27 we were building cars before we had a labor force we didn't have service stations and part stations we didn't have roads and infrastructure for them to drive on we certainly didn't have driver's license and auto insurance requirements we've evolved and changed yeah we when an industry becomes mainstream it adopts regulation governance some sort of oversight and ethics as a byproduct of that those are the things that our industry is striving for now we've self-imposed to groups like infrastructure masons and groups like that we've come on and we've created these self-imposed accountability groups that said hey these should be the standards and we should hold ourselves to a higher standard than the public would anyway yeah right because we know that they're afraid that we're stealing their energy what they need to understand is we're no longer data centers are becoming utilities they're no longer just consumers of the utility grid they are producers and partners to the utility grid i mean 55 gigawatts of data centers online today knowing that we're tripling but for the end of the decade and simultaneously we'll add 40 to 80 gigawatts of power generation behind the meter which will reinforce the infrastructure of the grid if we have to design it in a closed transition we can co-generate back to the grid and move the needle to the left [00:15:38] Dan Crosby: so we need to go and figure out what heavy industrial has done for a long time so i mean right now the [00:15:44] Kirk: number one consumer of energy in the united states is the cities and the homes within them right i mean 75 of the water that's used on this earth right now is coming from agriculture and that doesn't include golf courses which are a larger consumer of water than a data center right and like i said we don't use water in data centers anymore for the same reason why you don't use water in a closed loop system or a radiator in your car it can't absorb and reject heat fast enough and we and we run with gpus and cpus that are so dense now that we need to use some sort of mineral fluid glycol or a dialectic of some kind that can absorb and reject heat faster transfer so that's why we don't need water i'm not saying when we did evaporative cooling we weren't using a bucket of what we were using tons of water but now you know we build a data center that needs the same amount of energy as a city not only can we build that energy for the city but we don't use any of the water that the city would require well [00:16:37] Dan Crosby: and that whole thermodynamic equation is being worked through right at the end of the day finding work for that heat and not just getting rid of it is going to be the next thing that you see to start really and if you think about like what was driving all of these efficiencies of being super efficient with energy being super efficient with water is economics you know when that's a massive input into something the more that you can do something to save it the more you're going to reduce your operating costs the more profitable it's going to be and now you know for a lot of these facilities especially if you're trying to use legacy the denser and more efficient it can be the denser it can be um and so that's you know really what you're we're seeing that power density shift christian blotty [00:17:17] Kirk: was on and he uh was one of the biggest brains in our industry still is a massive influencer to industry but he came in my he was from microsoft and and he tells a story on his podcast with us on how he's coming back from japan and he comes up with the pue formula yep right so in essence for those of you that were listening data centers were just massively growing in demand of energy but if you looked at it from a percentage of baseline load through the grid it maintained a point two to two and a half percent utilization rate for several years more than a decade and why is that it's because the pue made it to where we were now measuring how we drew more efficiencies right out of the data center so it wasn't that the demand the demand was skyrocketing but the grid wasn't seeing it relative to the other demands of energy because as we demanded more energy we also got more out of it and because our utilization efficiency got so good it looked like it was flat for years then all of a sudden we almost hit a plateau on innovation and now you can start seeing the real linear trend that we're taking place now in energy and let me zoom us back out to where we started with this one particular report because yeah i want people to understand you know what what does demand queue look like and you know they use phrases like speculative excess and this and that and what's the net what is the u.s grid right now what is the in the queue for the u.s grid let alone [00:18:43] Dan Crosby: just texas so this isn't and we put i actually put this up on the screen at 2044 dc ac yeah i remember there's two different queues in a power grid so there's the generation q and then the load queue the generation q has been jammed and speculative forever and the reason for that is if you're going to build a generation project you need to figure out if it's going to be economic you need to say [00:19:05] Kirk: forever before data centers oh yeah what is the largest demand source for energy on the grid outside of residential today heavy industrial heavy industrial and they're way bigger than the demands of energy for [00:19:18] Dan Crosby: data centers today am i right well the data centers are catching up and exceed i mean if you look at five years ago you know big chemical plants and those kind of things which were in the hundreds of megawatts um were kind of the biggest loads out there and that's where there was this you know really probably back in 21 is where we saw this you know leapfrog of there aren't going to be just a couple of these isolated you know very big hyperscale data centers but if we're going to you know go after ai and the way that we're talking about the density of this is going to skyrocket and so that's where in a lot of ways data centers have jumped to the top of you know size wise of the industrial market but if you also think about it it's a very different process so there's so much more control and you know how that load can be managed is dynamically different than a chemical process that you know has a whole bunch of momentum to it and other things that you can't just turn off and stop or oil and gas where if you turn off a well you can kill it so you know this is a completely different scale of load that we've ever seen before but it's also a different controllability of a load that we've we've ever seen before and so there's a couple of things to unpack with cues right so if i'm going to build a solar plant or a new power generation facility i need to figure out how much it's going to cost me to interconnect to the grid right because that's going to be a big part of tying into that market and then i need to go raise capital and do all these other things so they would go to a grid say you know i need to do a study what is this going to impact and especially with renewables who have you know solar has a relatively predictable shape but wind is you know pretty dynamic as far as when it runs and so the grid has to do a study so you will get these huge queues and forever we've watched this because you're looking at the supply demand balance of a market and always have been and a lot of what's in a queue you've always discounted is it probably won't show up because a lot of that is economic research and people spending money to see if a project works and then it doesn't and but on the load side you really didn't have that right i mean unless you were going to actually build a facility i mean you went to a utility of about as much excitement as you went to the dmv right um and so people weren't just showing up at the utility unless they had a project that they were constructing and most of the time they've already done all of the financial analysis of i'm going to build this chemical plant or an lng facility or a you know pharmaceutical manufacturing whatever and so coming to the utility is just part of connecting that project and turning it on and so that project that process was very straightforward um you know there are definitely complexities to it and there's usually been some kind of customer in aid of construction and you know any facility that was actually serving more than one customer got socialized into the grid because that was going to get recovered through the tariffs and and that's the way that's always functioned and so if you look going back to like you know i mean we were talking to people about data centers in 22 and really 21 and you know people what's the data center you know what's the transmission line how does this work like no one was you know involved and now it's funny i from a family you know forever ago people would ask what i do and you explain it they're like and now they're like oh oh really and it's because everybody now knows what a data center is and a transmission line and a gas pipeline and a power generator an average data center is now no less [00:22:26] Kirk: than 200 megawatts right right and back when we were doing big deals when i say big deals i mean 10 20 maybe even 30 megs in a building was huge it wasn't so disruptive to the utility provider right [00:22:37] Dan Crosby: well those are distribution loads that sometimes were actually fed all the way back to substations redundantly but oftentimes came off of distribution services because that's a load that can be served by a distribution service once you move you know above 100 megawatts you're moving out of sub transmission and you're moving into transmission service that's just the way the poles and wires work and so the big shift here was you all of a sudden have all of this capital flowing in this direction and everybody that was even change tangentially related to real estate was all of a sudden a you know data center developed land powered land developer everybody that was a reason or you know real estate developer all of a sudden was a powered land developer and so there was this literal gold rush of everybody that there's a transmission line there so i have power which is not how it works um and all of these projects you know start materializing out of nowhere and quite frankly it really caught a lot of utilities off guard and it was interesting having you know worked with so many utilities in the power and gas side for a really long time we certainly saw on the power side first and then the gas side and you know many people i've known for 10 15 years were like what all these people are calling me they have no idea what they're talking about and asking for just you know absurd amounts of load and not even really understanding the weight or complexity of what they're asking for and so you developed this huge queue and and from the outset i think everybody utilities and anybody in energy knew that most of this stuff was exactly that it was total speculation and people just trying to get a load letter from a utility um and the issue was is is from the beginning of the process utilities have always kind of taken people's word for okay well you're coming to me and talking to me this must be real you wouldn't be here for fun and very quickly um you saw that that they are here just speculate and we saw a bunch of this in bitcoin that the thing with bitcoin though is you could stop it with money you could just be i could cost a hundred thousand dollars for that study totally right and and like a lot of utilities like encore set up a separate queue for crypto and everything was paid up front because they knew a lot of it was speculating yeah so all of a sudden you have big capital that has no problem writing a million dollar check for a study and we saw all sorts of people from all walks of life that had money and you know this was the new lottery and your ticket cost a million bucks and you could get 200 million back um and so it took a little while for the the utilities to really catch up on raising the cost of those um studies and and what's interesting too that's going on right now is every utility is handling this differently um across different parts of the country we've seen actually there was a utility in the midwest where they're still not charging you for a study but you're coming with full engineering geotech every you're ready to build yeah and then we'll talk to you um and because that's way more difficult than money right and which is you know an interesting way to go about it i'm not saying that's the right or wrong way um you're having utilities saying listen you pay for everything when you energize i'll pay you back the utility is saying you pay for everything i'm never paying you back and you're going to pay my tariff which i think will be a problem down the road um and now we're seeing letters of credit that are literally in the hundreds of millions of dollars pushing a billion um if you're getting into gigawatt scale projects in some markets and so every market is is working through this actively and a lot of that is really focused on if these projects break and 100 million dollars with the infrastructure has to get built that doesn't fall to the rate payer which is great um and that's the way that it should be and and one of the things that's changed too is even when we're working on like you know 10 years ago working on the lng project and huge 345 ring bus you know 50 or 60 million dollars worth of infrastructure but it's still not going that deep into the grid with some of these projects now i mean you're affecting utilities that are two utilities away i mean normally we would see you know four or five thermal voltage violations on a transmission study i saw one come back with 54. so you're impacting grids in the way that grids haven't been impacted and wisely grids are saying well wait a minute i don't need to do all this work if you don't exist um so this really is you and here's how we need to allocate this and there's a lot of things that are still getting figured out you know some of this i think is gonna end up in court um but that's all part of figuring this out and transitioning you know through this process and so you know talking about the queue being catastrophic and the queue you know wrecking things i think the biggest thing that the the biggest frustration in my opinion out of the queue is you have a whole lot of people that are effectively just trying to get in between the data center builders and the hyperscalers um and see it as a market opportunity right and that's obviously very frustrating for the data center builders and the hyperscalers and but when you start putting up these kind of you know think about how many entities out there can write a 700 billion dollar letter of credit and so if you think about what the utility is trying to do is it's they're weeding now let's weed out i know who actually can do this so here's what you need to do to do this and and then we'll go um and for the people that want to you know pony up and pay that kind of money for a risk okay fine but that's a huge risk and there aren't a whole lot of people willing to take that and and that's where i think a lot of you know projects are starting to get jammed up as they're they're realizing they need a lease like in northern virginia they're like you're [00:27:56] Kirk: out of power right and it's like we're gonna claw back what you're not using and there's groups now we have a liquid cooled load bank business called patriot power services that doug runs and right now he's taking inquiries from people that are like hey um we have data center capacity in california is an example and if we don't utilize more of it and they're going to claw it back but we have contracts with our customers that haven't ramped to it yet so they're asking us to put load banks on utility just to simulate loads so they don't have their energy cloud back what i'm telling you is like we're seeing some really weird things happening because people are protecting their [00:28:27] Dan Crosby: energy and that's not fortunately or unfortunately that's not that weird um and and what's interesting is it utility structures exist for a reason right and the utility is there to make money and make a return and the flatter that that load is the more efficient that utility is which is why they want that capacity utilized and whether it's utilized by you or somebody else um fine the other thing that's been going on forever is optimizing capacity and transmission tags and we have you know big industrials that will help manage those loads during peak periods and you can look at it as it's gaming the system but not really right i mean the reason that those capacity tags are set when they're set is this is what your contribution is to the peak so what i'm stressed out as the grid and trying to come up with as much power as i can because my primary you know person i'm going to serve as a residential customer you're going to pay for that well if you disappear and aren't banging up my peak and contributing but you're actually helping me okay well you're not going to get hit as hard and so and i think this is going to be a big part of what happens next and there's a ton of stuff going on with nurc and a bunch of the grids right now because there's a lot of physical dynamics that are going to need to get managed here but the benefit is there's no load out there that is you know maybe crypto but other than crypto there's not really any load out there that can dynamically manage their own load the way a data center has and a lot of that comes from you know needing nine nines yeah um is that you've you've got to have that ability to function even if the grid can't yeah go to island mode data centers are designed to be in island mode at all times right and that's going to be an emissions challenge of you know in different areas of being able to run that backup gen you know more often than not some of this is going to be on-site gen and net exporting to the grid but the whole concept of like data centers are going to show up and steal my power it's not how the grid works right there when when the grid gets tight there is a curtailment order and it's going to start with in heavy industrial aka data center and it's going to end with hospitals yeah that's how it works so it's not like data centers show up and and steal power and and the dynamics of power markets are electricity is the most volatile commodity on earth it can't be stored it's kind of nerdy but you don't store electricity in a battery electricity is kinetic you convert it to chemical energy or whatever that battery is using to store and then you convert it back to electricity and you have losses on both sides of that conversion right so all of these power markets have to be managed instantaneously and too much power rolling blackouts not enough power rolling blackouts and you're maintaining a frequency of 60 hertz which has to be very dynamically balanced which is one of the things that that nurk and a bunch of the grids are really diving into right now is the the real [00:31:14] Kirk: complexity of that with are we more stable base load compared to traditional industrial manufacturing that is cycling up cycling down for sure today i mean are there favorable advantages to our our energy [00:31:25] Dan Crosby: demand that make it more stable for the grid yes and no so i mean it depends on what the ai load looks like and this is a lot of what people are talking about training right is you mean you're you're federating computers that can move in ways that quite frankly only an arc furnace does in the grid right now there's a lot of protections that we put in place with an arc furnace and you're saying the same thing now with data centers and how that they're you know low voltage ride through and a bunch of these things that we're working on on a technical level which are going to have very big impacts on markets but as far as being able to respond there aren't many loads out there that can respond as quickly as [00:32:00] Kirk: quickly as you drop a data center in a heartbeat and they'll run on engine without seeing the lights [00:32:04] Dan Crosby: flicker in a literal heartbeat yeah and here's the thing dropping or adding in a heartbeat in and of itself isn't great for the grid so part of what we're talking about is okay well we don't want you to add or drop in a heartbeat we want you to do it in 25 heartbeats right um and so and but the beauty is a data center can do that yeah um and so building a lot of those dynamics and and this isn't there's there's two pieces here right one is the physical you know managing voltage and harmonics and frequency and the other is managing the market forces that go along with that and that's the other part of electricity there's there's no commodity on earth that can go from negative twenty dollars to five thousand dollars in 15 minutes only electricity can do that and so but those dynamics and those incentives are there for that reason is if the market goes to five thousand dollars and i'm gonna have to spend five thousand dollars to stay online i'm gone and that's why those dynamics are are set the way they are and all of these markets have guardrails and caps and certain things that they can and cannot do financially um to keep that you know constraint and and the difference in those grids is wild um even across the united states so a lot of the com comments about you know data centers are are making you know electricity prices go up kind of going back to causation correlation it's just not true as of right now now could some of those things happen in the future sure but this is all part of how markets respond um and we've watched this happen you know we went through coal to gas switching and all these other market dynamics over the last 15 years where you know a huge percentage of our fuel mix has completely changed um in the united states and so there's there's way more going on than you know data centers are are [00:33:48] Kirk: driving up power prices well listen man real quick just to dovetail into that the just for our listeners again once you're listening if you haven't tuned out yet the reason why i challenge you to listen is we need to we need to understand this we have to protect our jobs if we don't understand this and we let nimby protesters shut this industry down then those jobs they go overseas so let me just pause on that part i mean we in the late you know between the mid 80s and the early 90s we were gutting the entire middle class of america because we outsource manufacturing now the most important thing to manufacture is an ai factory and for every ai factory we build we have to build essentially eight other factories that build components that we could use to support that gigawatt factory that ai factory we have to understand how to express and explain these things at cocktail parties at happy hours on the golf course and just sitting around with other people that don't know the industry because it's wired in the tenants of our human biology if we do not know something we fear it and right now people have a lot of fear over what this industry offers and i want to reshape it or reframe the way that we view date don't think of an ai factory as a building look at it as a collection of humans and those humans are all getting new meaningful careers with a high overall total learning potential that doesn't require them to go to college and they could have meaningful jobs that have a significant impact on the way that we i mean i challenge anybody to tell me of any type of technology that i have a greater impact on humanity since powered flight yeah right so we are working on this is the fifth industrial revolution and the third was when we were getting technology at a mainstream scale and maybe that was a little too hot and then we over rotated on technology and the fourth industrial revolution was a little too cold and the fifth industrial revolution is intentionally designed for us to have a healthier more harmonious relationship machines and technology in our life that is the goldilocks moment of technology and we are we are spearheading it we collectively need to be able to defend this industry and with regards to energy i mean from a q perspective texas isn't an outlier to the country right where we have texas well texas is texas yeah well texas is texas but how many gigawatts are in the queue right now across the country [00:35:57] Dan Crosby: would you estimate hundreds and hundreds i mean there's over 200 i think right now thousands but everybody knows thousand yeah but everybody knows an overwhelming majority of that is just not real um as that's where a lot of speculation studies are coming from yeah but so i think there's a couple of really important things to unpack in what you just said right is is there are things to be afraid of an ai one of them is losing the ai race yes so um at the bottom level not only that so people [00:36:24] Kirk: understand losing the ai race means not from just a national security perspective where now some other country that doesn't share our beliefs or our common um uh commitment to fellow man um those countries that they are going to most likely end up building in are going to be massively deregulated when it comes to environmental standards so if you want to see that in china right yeah how many and they're still building coal two-thirds of the world's population sits in asia specifically between india and china and that represents almost 80 it represents 75 of the world's manufacturing almost 80 of the world's pollution so we have the ability to not only build the most efficient buildings on this earth but we could manufacture our way through this problem while simultaneously reducing the impact on the environment because if we outsource it to another country we lose another job well there's no outsourcing ai i mean whoever national security defense is going to require us and that goes back to the circular financing concerns that people have which is also followed in this report we'll get to but i think [00:37:26] Dan Crosby: important thing too and what you just said is is this is it's causing demand in the trades that we haven't seen in decades right um and so it's a huge opportunity of you know one for people in in the trade space to make i mean you're seeing electricians and trades people making money that a lot of people in wall street are trying to make yes um and then two there's that entire manufacturing supply chain which a lot of that for a long time in this space specifically especially transforms and that kind of stuff a huge chunk of that has come from overseas tariffs are having an impact on that where more of that will get manufactured domestically so there's you know the opportunity here is massive um for the united states and so understanding you know the benefits and listen there are scary parts and that needs to [00:38:14] Kirk: get managed and they are valid concerns right i'm not trivializing people's concerns on energy safety overall the impact on the environment how they're going to steal their water all valid but all addressable right and they're measuring it'd be like measuring today's automobile versus what we started with and they are not even close to the same types of products anymore yeah and it's because we were governed right uh the federal government came in and helped build scalable workforces to build infrastructure the groups that benefit the most are going to be supply chain groups and service companies those two will benefit the greatest from these things because as demand for data centers grows demand for supplies will grow which draws more factories henry drives innovation too 100 and like the beauty about like when i said henry ford not only built the automobile he built an ecosystem of a million small businesses that were driving that innovation and he was using it to transport goods and human beings now we're using ai factories to transport data and app services those app services like cloud-based tools are what give small businesses an advantage to compete against the large businesses because i don't think any of us want an oligopoly where there's only five to ten companies that are making a ton of money and controlling the entire way we build these things look at some of the stuff that [00:39:32] Dan Crosby: people are able to create in 10 minutes that used to take you know 10 million dollars in two years and [00:39:38] Kirk: ai's level in the playing field you don't have to be the biggest company or the wealthiest company to compete you can use the same ai resource tools that they're doing and it puts the small business right on the same playing field which is why today we represent 75 of the workforce so you know i'm trying to draw this back because i need people to understand this is about humans right ai the ai boom is not about technology it is about careers and changing people's lives because we will start manufacturing the most important part of technology ever built and ai is the most advanced weapon machine and tech ever built it'll probably be harnessed and used first for national security energy second and health care third dynamite those are the top three reasons why we need this tool right and we can't relinquish our lead to that to someone who may not share our values globally right so we need if we don't solve for this as a collective industry then you will see jobs outsourced and that's what gutted the middle class are in the first place we are using ai factories as the opportunity to build back the middle class so this is a huge opportunity for the middle class to get [00:40:43] Dan Crosby: involved um through the trades and and that's and i think some of the stuff that you're working on right now setting up some of those programs to get people into the trades you know pretty quickly and you know when there's this volume of work there's a tremendous opportunity to learn and that's where i mean you saw a lot of trades that started to die because there wasn't a ton of work and so a lot of those skills weren't getting passed down and we've even seen this in other industries like an oil and gas where you had literally a generation gap um in that industry that's been moving its way through because everybody went to a different space and so um you know that's another you know really big opportunity but dispelling the myths around a lot of what's going on in power markets i think is [00:41:23] Kirk: the doom and gloom this thing was pretty hey man the sky is falling and everyone's lying right about what the needs are and we're like we're not lying everyone's kind of like many people get reservations to go to a restaurant and then they back out at the last minute but at least they want to hold their spot in line in case they choose to do that some groups will do this time they're holding their spot in line and then trying to figure out how to make a pencil out and if they can't then they abandon but [00:41:47] Dan Crosby: this if you think about like let's talk about the rate payer right because at the end of the day the whole purpose of these concerns is the rate payer the biggest person or the biggest group that is being impacted by the queues is the data center builders and the hyperscalers yeah is they're being slowed down by it because it is gumming up the works but it's not you know gumming up the works for someone connecting their house to a distribution power line like that it's just not so what's happening i think [00:42:13] Kirk: the reason why it's gumming the works is because the demand in which we build and scale at the volume and velocity that we build at now is driven by the the adoption rate of emerging technology which moves really quickly but energy moves at this at a political speed which is measured in decades so that's why it's gumming it up we move at the speed of technology and the adoption rate of it they move in decades we move in i mean this industry and part of it though i think there's been a rude [00:42:41] Dan Crosby: awakening for a lot of people in the data center space of understanding just how long transmission power has always taken it's killing how long generation has always been and really you know it's not like building a transmission line went from six months to five years all of a sudden you know these projects have always taken a long time with ccn's and eminent domain and sourcing all of the equipment tons of this is just completely custom manufactured and you know things that used to take 18 months are now taking 24 36 things that used to take 24 to 36 are now taking four to five years and part of that's because of the queue and again there's tons of speculation there there's a whole bunch of p funds that are buying transformer slots and buying generator slots and trying to get into that supply chain too and making margin um and so you know a lot of that is what's going on and so i think there's there's kind of two major things to understand here one is the cues and it's not just cues to get into transmission it's cues for transformers it's cues for everything that's in the transmission class and then there's the the whole market side of what is this actually doing to markets and and i think there's a lot of it's all about headlines and we talked about this actually at dcac back in 24 of everybody putting up the pjm headline of you know capacity went up 9x okay fine then we put the chart up there it's it's only barely double what it was six years ago right but capacity had gotten cheaper and cheaper and cheaper and energy in general has gotten cheaper and cheaper and cheaper and one of the things we showed like in the west texas grid you know power prices are relatively close to where they were a decade ago okay so data centers are not driving up right but but there's also been times where they were you know 100 times higher um like in the case of yuri of where those power markets can blow out because of supply constraint and if you think about all of the things that have been happening over the last 10 or 15 years in our grid right i mean renewable portfolio standards the proliferation first of wind across the us which is very difficult to balance i mean alberta canada when wind gets on the margin of your power grid you're going to be borderline free or at your cap there really isn't a whole lot of middle ground there and so as a industrial user trying to come up with a budget in that kind of market all right well tell me when the wind blows and i'll build you a budget right so so that's challenging but it also has tremendous amount of benefits and in a lot of other ways and so you have a massive influx in renewables a lot of it was funded by itc so it didn't really fall directly to the ratepayers it fell to the taxpayers but also totally changed the way that grids had to be balanced and as we saw in spain inertia matters and so being able to have you know traditional spinning generation sources that aren't using inverters to basically mimic the frequency of the grid is really important and so balancing grids became much more complicated and then you know you have a whole bunch of other factors like what went on in ukraine i mean at the end of 2022 gas prices the united states about the nine dollars you know we're three now so and we've seen crazy you know short-term spikes there's sometimes the northeast the gas goes to 120 a decatherm but we saw monthly settles that were getting up toward nine bucks and part of that was the war in ukraine right europe all of a sudden stops buying tons of gas from russia goes into total chaos and if you look at all of our lmg export terminals i can sell gas in europe for like literally 50 a decatherm so if i'm here in the u.s buying gas to fill it i don't care what the price is just give it to me yeah and so that's limit up but there's also only so much we can export every day so once you hit the cap that's all the limit up buyers that there are and we're still constrained right and and even now i'm gas we see gas in west texas trade negative where literally the producers are paying to put it in the pipeline and you're seeing some data centers trying to go out there and build and we'll see you know how that works out a lot of pipelines are being built to bring that gas to other markets you know new england has been a super dynamic market for a long time in fact it's the only market in the u.s still where oil is actually on the margin if you look at some of these gas power plants they have huge tanks of diesel that when prices you know and this again this is the dynamics right is if it's 15 degrees outside i want that gas molecule going to grandma's house and not to a power plant so the gas price explodes grandma doesn't really feel that because she's on a residential rate well now that gas plant can't afford gas so they're going to burn their diesel to make power and so all of these mechanics are are there for a reason we've also seen a huge rise of inflation so if you look at like 11 of the top 15 data center markets in the united states have seen around average or below average um rise in electricity prices over the last five years and if you look at you know northern virginia for example middle of the pack i think they're like 28 percent you know maine went up 60 something percent california went up 60 something california had to you know put 27 billion dollars in fire damage into their tariff so even when you look at how these different markets they all function completely differently in different ways some have capacity markets some don't so to try to associate a rise in prices across the u.s one that's hard to do um and then two to try to allocate it to a specific industry again is really hard um and especially when you think about you know the volume of build out really hasn't materially hit the grid yet it's just starting you mean behind the meter well and and the volume of data centers getting built right yeah i mean it just takes time 10 gigs online this year right so it takes time to build you know people talk about hundreds of gigawatts right it's it's been in the 10 about range right and and if you look at a lot of the utilization of grids in these markets you know a lot of grids weren't utilized which is why capacity prices got so cheap now we're seeing capacity prices clear at the cap well why is that if the price clears at the cap it incentivizes generators to build look what we did in texas right we put the tef in place which was to incentivize more thermal generation right because thermal generation has inertia it's much more stable and it's totally dispatchable i can call on it whenever i want i'm not you know it doesn't need to be sunny or windy or or a battery can run until it doesn't have any power so we incentivize them in texas with this really really cheap industry interest rate will finance it but you have to export at least 50 percent of what you generate well the problem is power is so cheap in texas it's not economic to run that plant dispatcher like on our platform i literally get texts every time power prices and texts they have go negative you're getting paid to use electricity which happens all the time especially in the west zone because of renewable generation and so right now there's a lot of markets where it's not economic to generate electricity so if you look back historically historically prices are pretty cheap i mean even in a lot of these markets we're seeing prices that aren't higher than they were 15 years ago forget inflation like dollar for dollar and so so that's one of the when you see this rise over the last couple of years especially if you look back about three years before a lot of this started to come into place prices were in the tank you couldn't build a gas plant in texas it wasn't you couldn't afford it and now what are we seeing you're [00:50:04] Kirk: seeing groups like caruso at stargate that are building their own power behind the meter entirely well [00:50:08] Dan Crosby: you're going to have to um and there's two reasons to do that uh one is just on a macro level right if we're going to put this much power online you have to build more generation there's just not enough and that's why you're seeing you know china's building generation out the wazoo right now they [00:50:24] Kirk: have commissioned nuclear reactors every year for the last four years right two year except two years ago they did they did 10 every year with the exception of two years ago they did 11. i don't know how many of they finished out this year plus to my knowledge they built 100 coal power plants last year as well so they're going to beat our ass on the power game and these things these ai factories just like every transformational and industrial revolution we've had the number one fear has always been what are we going to do for energy and how is it going to impact the environment it's no different than what we're doing right now so those people that are listening what dan's talking about right now about ai you should interpret into your ears as opportunities for new careers because building energy and our head-on collision with the the only reason why we are no longer operating in the shadows is because the demand of energy has increased because the size of data centers have increased from average of 200 megawatt and a minimum to one gig and the density has gone through the roof intensity has gone through the roof so deploying large volumes of aggregate technology with a huge low demand of energy with a very limited amount of water demand now is the new norm these are the most efficient things ever built and not only that although i said we're going to triple in an industry size between the end of the decade everything that we built to date will have at some point need to be retrofitted back to an ai data center capacity workload which will make those ones more efficient yeah so we're going to continue to drive efficiency back into these facilities which will continue to rely less on it takes a greater risk away from a consumer who's worried that we're going to build a data center that's going to take their energy and their water they need to see it as a place that we're going to build for them that's going to build more result reliability under their grid and offer them a meaningful career and opportunity because we're going to be hiring locally where we're building these things here's the other [00:52:10] Dan Crosby: thing that's i think really important people to remember too is and we showed this i showed this chart is like 1955 to like 2005 you know energy usage in the united states was up into the right and it for 20 years was flat and if you think about what was happening efficiency was gobbling up load growth but two when you have 20 years of stagnation in the market you're not able to put money into infrastructure because every dollar you put in infrastructure is going to go to the rate base and if you're upgrading things that you think need to be upgraded but might not have to be upgraded their rate base because whoa why are you spending that money and making me pay for it so if you think about what's actually happening here um a lot of these grids are very very aging um especially on the transmission side you know some of the stuff has been in place for a very very long time like 10 12 years behind on transmission right exactly and and so i mean and i think i put that chart up there where the only major transmission project we've had in the last 10 years has been the 20 years has been the crez project in texas and that was really to move all of that wind generation from west texas was trapped and literally driving prices negative every single night over to the rest of the grid where it could actually get utilized and so if you look at what's happening here billions and billions and billions of dollars are getting put into infrastructure to upgrade it and make the backbone of the grid far more resilient and far stronger than it's ever been before and it's being financed on the back of these data center companies and that's part of saying hey yes i need to upgrade this transmission line that's you know literally 70 miles away from here but i wouldn't have to do that if you weren't putting 600 megawatts over here so you're going to pay for that and it's like all right fine and so that's now getting paid for and so do you think about the reinforcing that does for i mean that line technically serves hundreds of thousands of customers potentially so the the ability to really upgrade and transform the power grid and the dynamics of our here and if you think about what that does on an ai perspective of all the things that we're going to learn and figure out through that and so that's the the other piece of this is the other reason that a lot of behind the meter generation is going to happen is because you have to physically manage and keep the grid stable and listen the data center wants the grid just as stable as the residential consumer right and so keeping voltage stable is going to be a really really important thing and this is what low voltage ride through and a bunch of things that the grids and nurk are working through is when you have these kind of dynamics they've got to be managed and and we're really pressing the forefront of what that can physically look like and you're talking about things that are happening intra-second um to be able to manage that and understanding how those things are going to physically interact with grids and how are they going to financially interact with markets is all part of what is being worked through very actively right now it's very complicated to [00:55:02] Kirk: construct those two things but i do see a pattern which and you know i don't speak on behalf of any of the hyperscalers but i've seen uh some of the reports i'm sure you've seen them too where they're already going out there and talking about how they're going to build their own power generation and it like i think people need to be aware that we're already doing that we're building tons of power generation like i said before the end of the decade we'll put 40 to 80 gigs online behind the meter as a data center industry alone yeah so we are partners and not just consumers of the grid we actually build infrastructure that reinforces the reliability of the grid and it's with that will come a ton of jobs specifically on the supply chain side and the services that follow right so there's going to be a ton of tax benefits too for every 100 billion dollars of infrastructure built on data centers it generates immediately 800 million dollars of tax revenue that equates to 40 million 40 billion dollars of labor revenue which could also be taxed and not to mention what they spend their money on from a sales tax basis in addition to that that puts a bump of 140 billion dollars into the gdp yeah we're on track to spend five trillion dollars in the next five years in this industry so there's so much of an opportunity for new careers new jobs going going back to uh that the national grid i know that we're texans so we're going to talk a lot about texas and i think the sky for the cloud will always be northern virginia but the home for ai is going to be texas though not to say there won't be ai factories in other states but what's driving them to those other states it's we're going to where the power's at [00:56:32] Dan Crosby: it's pockets of power um so you know short-term power availability has a massive premium on it right now and quite frankly most of that game is played out uh and this is why you've seen a lot of data centers go to you know really small markets that might have had you know incremental capacity a lot of it was from manufacturing that left so again if you think about what this does when you have you [00:56:55] Kirk: know like in the rust belt where you have huge distressed cities that were gutted because we outsourced the manufacturing there they still have power plants and manufacturing plants that we go build [00:57:04] Dan Crosby: data centers right and a lot of that power infrastructure some of it's still there and being reutilized um in fact uh you know i've seen one that was literally using an old transformer from a generator that had been decommissioned that they were able to use to get the facility online faster until the new transformer showed up so you know in a lot of those markets that's why you've seen significant proliferation now most of that is disappearing and one of the challenges too is when you have loads of this size i'm dealing with one right now where you know we're bringing a load that is literally a third of that entire grid's power so you know they haven't managed that kind of load before and that's where you can really partner and help them understand how it's going to work and work through it together of how to make sure one that doesn't fall down to the rate payers two there's a benefit to the rate payers and then helping them actually more strategically manage how they were managing energy to begin with um and so there's there's a lot of grids where even the load might be bigger than that grid and quite frankly what we're talking about here is bigger than a lot of grids even in in texas and pjm and some of these other markets if all this is real but there's also another component too of what does this geographically look like it's going to come back to fiber it's going to come back to connectivity to eyeballs some of this stuff can certainly be in the middle of nowhere a lot of it needs to be close to users i've seen actually a pretty significant volume in fact i had a call this morning of a lot of these 30 40 megawatt data centers in cities really starting to um ramp up that's ramp up uh because they need to be you know really really close to users for latency so it'll be interesting to see how all that um plays out well that's the demand [00:58:38] Kirk: driven from the inference side of the data center or the ai data center these ones that we're building out in these communities that are either distressed because they were once manufacturing cities that industry was outsourced and the jobs went with them or new markets all entirely were going in and we're building you know today we're being asked for you know 100 people at a time and and and we're being asked for the the man camps that would go with that labor right so it's a completely different monster the volume and velocity of scale now the only thing that's holding us back from adopting if you look at the the utilities as a general the very first one was plumbing and we created plumbers and in the early 1890s we were all chasing illumination i mean so we were chasing the light bulb and that was the generous of electricity we electrified the grid we went through the current wars between what would we use to standardize as a grid would be ac or dc power and we created in 1890 we i mean there was no there were craftsmen that were working on wood or working on metal or fabrication of some kind and we were recruiting them into this program and we called them electricians and then by 1894 1897 the ibew was formed to standardize on what we trained in curriculum we were creating trades as the utilities were being built yeah the first was plumbing plumbers second electricity we had third which is connectivity thank you ma bell for putting a copper line in our homes right the fourth was gas when you add those four together you have the fifth utility which is cloud which is ai and we use it every day we need to work with the department of education and the department of labor to work on standardizing data center apprentices and data center journeymen and and creating a trade that is more homogenous to what this industry really truly needs demands and we can't send them to college for those degrees by the time whatever they start studying their first year in college on data centers is obsolete by the time they graduate four years later so we just need programs that are like the trades and and we're working towards figuring out how to solve the labor side of this industry because when we solve for labor we'll solve for energy because the energy is going to come back to you're going to have to manage the capacity capabilities from a technical basis on how you interact with the grid at the same time you have to measure that simultaneously with what the market demand is willing to pay because everything begins and ends in finance it seems like right so while you do that there's going to be shifts in the adoption of generation of energy i believe that a lot of these hyperscale operators are turning into utilities themselves so that's the island mode side where they're building their own power plants behind the meter that can reinforce the grid but what does that mean to the consumer and is there any states i mean do you have any data that you could throw at us that talks about you know hey uh you're not going to see it in this state because it's over regulated or they just don't have a basin of energy or they just don't have whatever we need but these are the markets you're going to see a [01:01:21] Dan Crosby: consolidation into yeah well i think you know the northeast has definitely been probably the most challenging area and that northeast has always been a very challenging market for energy regulation in general is that what's regulation and also the gas infrastructure in there is just very very constrained so there's a lot of utilities that have little lng plants to be able to manage stuff in the winter there's in fact entire pipeline and distribution systems that are propane and so that's always been a very constrained part of the country so i don't think you know that's why you're not hearing a lot of data centers up in new england yeah uh is it's just a very constrained energy market quite frankly it's why a lot of people are coming to texas you know texas is one of the most if not the most dynamic power market um the us pj i'm seeing a ton of load we're seeing some stuff going on miso and out west um and you know there's a whole confluence of things coming together uh one of which is you know renewable portfolio standards and what is that going to look like with some of this load growth but two the dispatchability of power becomes incredibly important when you have this kind of base load and two or three the the dynamics of what's going to come out of ai and how to very actively manage these loads and here's the thing if you think about a residential customer they pay way more than an industrial customer has and for a long time and the reason for that is again power has to be perfectly balanced every couple of seconds right and you as a homeowner want nothing to do with that and so you're going to use power and do whatever you want and if you think about it we've been electrifying everything right so the electrical intensity of office buildings and houses and all that stuff is is going up as electronics were added but also going down as efficiency happened right and so there's a whole confluence of factors but industrial customers have been dynamically managing load forever we have tons of customers that are writing index wide open or using block and index or heat rate products or shape products and we're very actively responding to what energy markets are doing because energy is usually one of their top three costs data centers are going to be no different and we've seen this forever in crypto i mean crypto is very actively participating in ancillary markets and responsive reserve the most dynamic parts of the energy market where you need to respond in seconds and so data centers are going to be doing the same thing here especially being able to do that with on-state generation and other technologies to be able to help manage the dynamics of the grid and that's what's you know the large load working groups and a bunch of these nerd groups that are working right now of really figuring out how all of that's going to function i mean traditionally when you connect an arc furnace the arc furnace is constantly communicating with the utility and the utility can call up say hey you're not running today because i'm you know got a plant down or i have a really high demand and and they're always on an interruptible tariff and so the same thing is going to happen here but you're gonna have to figure out how you're managing fuel you know managing this volume of diesel is totally different than managing you know five megawatts of diesel right so there's all sorts of dynamics that you really need to think from about from the infrastructure side when you first put something there and that's also the big difference with utilities i mean i can't tell you how many situations we've gone into where you know one people that have never dealt with utilities before and they're pounding their fist and i need i need a gigawatt and i need it in six months speed like decades yeah and the utilities look at them like you don't even know what you're asking for um and so a lot of times the answer is no we had a situation recently where a customer was you know asking for power from the grid and a pretty big operator and said you know all we can give you 75 megawatts um and so they were like that's not enough and so we actually put a whole plan together of moving transmission lines some of it's going to be temporary and shuffling around went put this whole plan in front of the utility and they're like oh well yeah no we can get 300 there if we do that and so understanding how the utility functions how the infrastructure functions how the market functions and how any on-site generation is going to interact with all that and now you have gas pipeline dynamics behind that that's really critical um and and the utilities are listen they're in the business of selling power and they want to put that infrastructure in place they want to grow their rate base but you've got to be able to do it in a way where you're speaking their language and you're communicating with them and and we're seeing a ton of that going on right now the hyperscalers are doing a lot of the data center builders are getting a lot smarter with energy and having energy teams in place a lot of them we work with listen an educated customer is a better customer and so but that collaboration between data center operators and and here's the other thing there's a the data center operator in a lot of these cases doesn't have anything to do with what's going to go on the data center that's the hyperscaler of the tenant so helping the the data center operator and the tenant communicate with the utility on how this is going to function and designing power products that are going to interact and here's all that's going to change right as we're having to respond to these markets and the way that these data centers are going to have to respond that's going to create new financial metrics and new financial uh opportunities and the way we're going to have to change and manage the grid physically from managing you know current voltage and harmonics and frequency all of that's going to have to become instantaneous and we're already seeing that with some of these things where they're literally putting in fiber lines and under frequency relays and the whole point of that is if the grid gets into trouble the first thing that snaps off is the data center right and it's not like we're going to design for that right and it's not we're going to call you and let you know it's gone and like literally less than a third of a second homeowner's never going to deal with that right and so again there's there's so much good going on and there's also a lot of complexity and listen ai is going to do some things we do not want it to do like that's just computers have done that dynamite did that cars have done that like technology that's part of technology right but the more you can stay in the forefront the more you can really be a part of the solution and here's the thing the more you understand the more details the more you can understand where it can go wrong and put those guardrails in place and that's a lot of what's being lost in the headlines and the fear-mongering and things that are just you know getting thrown out there as if this whole thing is pretty linear and [01:07:27] Kirk: simple it is the antithesis of that so like listen i want to keep the focus just on power what are some of the other misconceptions that people may have that that just to kind of summarize some of the things you're saying too while you're doing it but what are some of the other ones that i haven't addressed yet you know for me it's reliability to grid where i'm telling you like a lot of these operators developers or hyperscalers are not just coming in saying i can build my own but they'll reinforce the grid they'll go in there and do maintenance to it and and do the things that have to bring it up to the scale that they need because they want to service it right and well the beauty of this is [01:08:05] Dan Crosby: we've never seen anything like this in the power grid right and and a lot of it is being driven by the cost of these chips and the cost of what you know the the power and when i say power like the economic power of what this can do is it's making huge infrastructure projects not a rounding error but they're not like incrementally detrimental to the project where for a lot of industrial projects you know going and doing 300 million dollars in grid upgrades is like a non-starter that project's dead i gotta go find somewhere i can plug this in and i'm not gonna have to do any upgrades right so you know those opportunities are evaporating and part of you know what and this goes back to this conversation we had about the queue right the the grid doesn't have an obligation to a generator a a generator has an obligation to the grid and in you know that was one of the disparate things with renewable generation versus gas and thermal generation is if thermal generation says they're going to show up and they don't they get hammered financially you can't do that with a wind or solar operator because they depend on the weather and so that's been one of the frustrations in a lot of these grids and where grids can literally operate you know renewable can operate profitably negative because of investor tax credits which obviously you know thermal generators can't do that and and nuke bids in at zero because they just want to run so even if they run for free for a day they would do it now that never actually usually happens but um that's all part of how that bid stack works but reinforcing the grid is going to be part of what happens here and in a huge way [01:09:36] Kirk: which is no different than how we had to reinforce the infrastructure behind the automobile yeah we built state system and then we had to build bigger roads and better roads i mean we're continuously there's not a city in this united states you're going to drive through that doesn't have some sort of road upgrade taking place right now the investment into infrastructure will never end and that's a similar pattern to the railway lines until we slowed down the growth of that and it's no different than when we were electrifying the grid which is never ending right so that there'll be a never ending investment into reinforcing the grid it just depends on who's going to be paying for it and we're [01:10:12] Dan Crosby: electrifying everything at the same time and even the oil field right i mean all of your we're going to highline drilling you know submersible pumps a lot of things that used to run on diesel you know ran now run on electric and part of that is electric is so much more controllable and specifically motors have so much more torque it's why a tesla can smoke a gasoline engine vehicle no matter what it is off the line is just there's so much more torque and control in electricity and there's also a lot of conundrums like we saw something in in new york that was kind of interesting and new york city wanted to electrify everything and one of the things was we're going to get rid of the steam system there's an entire municipal steam system that you know literally provides the heat for huge office buildings well one of the things that they've realized is you know electrifying everything isn't always the answer i mean one of the things we saw were a lot of these vrf systems where you're now using electricity for heat and if you think about it from an emissions perspective if i move a molecule of gas all the way to that hotel or whatever and i put that into a 90 plus percent efficient boiler i'm getting almost all of the energy out of that molecule into my building and wasting almost none of it if i put that molecule into some kind of generator i'm losing half of it probably thermally or close to it then i'm going to lose more of it transmitting and now i'm you know not very efficiently using what's left or maybe i'm super efficiently using what's left but i only started out with a third so when you think about you know all it's not doesn't always make sense to electrify everything and that's part of what we're figuring out they're not getting rid of the steam system because if you tried to heat those businesses those buildings with electric there just isn't the infrastructure and there's no way to upgrade it so now the steam system's you know really making comeback they're finding some other ways to you know they're looking at using heat pumps and some other things for that uh base system but this is the whole thing is is this is not simple there isn't a one size fits all i think this is what a ton of these speculators are figuring out is a transmission line just isn't a transmission line right there's different cable sizes different voltages there's different constraints there's different things it's tied to so every single one of these sites is different and every single one of these sites is going to deal with its own complexities every one of these utilities has their own rate base and structures and their own generation stack and how they're managing balancing all this stuff and so it's a really really complex situation that you just can't summarize [01:12:36] Kirk: very well so the misconceptions that people have is just because we have a lot of power in the in the grid queue there's a lot of power in the grid queue in every state and it's a lot of speculative excess [01:12:50] Dan Crosby: um guesstimating and there's a ton of speculation in the generation q2 [01:12:54] Kirk: and it hasn't been for a long time so for people that are seeing these types of things we're not saying that we're going to build all those things because this guy's like it would be impossible we don't have the infrastructure of the labor the reality is we know it's not going to be possible but we're trying to figure out of those ones there will be people that pop out and those deals will be real for sure and those deals that are going to be real are going to require a lot of jobs it's going to build factories so that we have a labor force that can manufacture components to build those substations and the grid infrastructure as well as the data center and that's going to require labor not to mention all the people that do it and in the markets in which we leave labor behind we generate six jobs for every one person we leave behind after a build but it's also important to [01:13:36] Dan Crosby: remember that like even looking at capacity markets and and prices the whole reason prices go up is it to make it more economic to build more and it's all about a supply and demand balance when you think about during a hurricane disaster right a generator that sells for 200 bucks is now you know some guy's got in the back of his truck for a thousand yeah right but in everybody you know a lot of people freak out and say that but at the same time someone's not going to drive a generator from 12 hours away for 10 bucks right but now a generator is available you can power up whatever you needed to power up and if it's critical enough that a thousand bucks makes sense then you do it right so and this is why we have guardrails in these markets is they can't go to infinity all of them have caps all of them have guardrails but those caps are set at a place where listen if the price is here i'm incentivizing new generation to get built and i'm creating a clear financial path for that to get built to rebalance and how those are going to be distributed across rate classes and all that kind of stuff is all part of the regulatory framework i think that you'll eventually see [01:14:38] Kirk: a lot more ownership being taken place by the hyperscalers and the operator developers for sure we're going to come and just build their own power because they will be partners to the utility and partners to the grid they will do it in a way that they can close their energy companies i mean this [01:14:53] Dan Crosby: is the reason that like petrochemical plants part of us they needed a tremendous amount of heat but most of these guys have always built their own power plants and you're even seeing that in some of the regulation that comes came through we saw some utilities that changed the way that they were going to manage capacity costs well the way they can change capacity costs is it basically is 80 of what it would cost to build a power plant well once you get me to that point i'm just going to drop out of the queue and go ahead and generate power right so you've made the economics for me make sense and i think you're going to see you know more markets doing that of making the economic equation there you know make sense because if i can you know if i need this much power that now becomes one of if not the biggest cost in my business outside of you know chips labor and supply chain so and so that's going to make that behind the meter gen much more economic and now you control the maintenance you control the reliability you control the redundancy and you know really control your own destiny and now you're functioning as a benefit to the grid of being able to participate and that's where some of these markets are going to change where the ability to jump in and help the grid out either by pushing out your excess power or shutting load down to even be able to help even more will become you know very lucrative you look at a lot of these crypto they're making more money in the power market they're making crypto and everybody's talking about how crypto is really helping the grid listen crypto is two million dollars of bitcoin they'd be mining you know to the moon but because they have very dynamic load and because they're able to really step in and interact with the market and the financial metrics and incentives are there and the rate structure they're making a fortune yeah and that's [01:16:35] Kirk: not going to change well i think it's common it sounds like it's pretty common that you see these these grid requests it looks like three out of four in the state of texas uh it's coming just from data center people and those data center people are testing the market to see where they're going to be able to get access to and and they're willing to make the investments into the infrastructure to get it or build their own right and we're we're going to see those commonalities and and why that's important to us listeners is that equates to new jobs you know again supply chain manufacturers are going to expand and grow service companies are going to expand to grow and there's going to be so many jobs that come with this i think that what other large energy consuming states compete against [01:17:18] Dan Crosby: texas at that level i mean you're seeing a lot of competition the pjm market for sure which is the market that basically runs from chicago over to new jersey okay so that includes [01:17:29] Kirk: pennsylvania pennsylvania okay so you're saying milwaukee you're seeing i just saw a big press [01:17:34] Dan Crosby: release about regarding one in in wyoming yeah right so those well wyoming's in the weck milwaukee's going to be in in miso but uh so each grid is going to have to tackle this problem for sure and and there's also a lot of starting to have conversations around interconnections of the grid and the interesting thing uh we talked about this two years ago back when wind really started to proliferate and we saw you know tons of wind in nebraska and iowa whole columbia river gorge out in oregon started to get built out west texas went crazy oklahoma but there was a whole conversation about putting this big 765 grid together where you could move power from different parts of the country where you have too much power over here move it over there and you know we have three different interconnections united states to the western eastern and then urcott's its own island and a long time ago they talked about a project called i think was called tres amigos where they were going to connect all three grids um over in west texas and you do that through dc ties so they don't have to synchronize and who knows you know as this moves forward there's underneath all of this is physics and physics doesn't bend um it is what it is and so you know part of how we can balance things is also going to dynamically change as ai is able to design better right and and here's the thing too in some of this it's so fast a computer doesn't have time to think like only physics is going to work and that's because electricity moves near the speed of light and so you know it is very dangerous especially at high voltage and and how quickly it responds in certain situations is is is one of the things that's really being worked through right now is how to balance that and and make that work from a physical grid infrastructure perspective and from a financial markets perspective and and that is a lot of what the conversations are going on across all of the grids in the country as well as in [01:19:24] Kirk: washington dc well i want to bring this one down to a close here in a little bit and i know i asked you a lot of questions what are some of the things that i haven't asked you you think are worth sharing that people understand and maybe use this opportunity to recap some of the things like you said hey energy prices have been maintaining for the last 10 years and if anything you know there's this reliability that's coming as a byproduct or what what do we want these people to know that we're we're not we're going to always protect the residential yeah we're not driving up the cost of energy and we're not [01:19:56] Dan Crosby: taking away from them right now yeah i mean well i think one of the important things is that residential especially rate payers or voters um and so you know and we've seen this forever of you know rate classes and industrials trying to game things and um one of the things you just have to be really cognizant of right now is everybody has the ability to put out voluminous information and cite all sorts of things because chat tpt said so yeah um or grok or whatever you're using and so there's a ton of headlines there's a ton of chicken little um and there's also like you know just flowers and roses right and and it's neither right this is a very complicated problem um it's a very complicated opportunity and a lot of things are going to be very dynamic you know i think one of the things that's going to be really interesting to watch and continue to see is you know i remember when cold gas switching was a big deal where gas really started to compete with coal and if you think about that that was because of you know directional drilling and fracking really brought the cost of you know a couple guys and three million bucks can drill a well now or you know only super majors could do that 30 years ago so technology is continuing to shift i i as we talked about on the last podcast i totally believe nuke is coming and eventually vehicles will come with a reactor that's probably about this big that runs for the life of that vehicle and you know we probably won't even own cars anymore and again utilization rate right think about how much time your vehicle sits still where it could be running around town helping other people and doing other things and it's just more efficient right so everything's going to become more efficient um you know i don't think nukes here in three years um i'm also hearing interesting things about some fission projects that are going on but i think that still has a way to go you know the reality is if we want firm dispatchable power there's only one place that comes from right now and that's natural gas at the same time there's an entire physical supply chain that is going to constrain this you can only build so many components and you only have so many people to install them and i think some of that will ramp up some of it is still skeptical of how far and how fast this is going to continue to go um but you know i think there is absolutely caution and caution is being taken and taken very seriously by the different regulatory agencies that are involved in this and we're getting the opportunity to have a front row seat and be helpful there just as we've seen a lot um but two is understanding you know where can this go wrong and how do we put the guardrails in place to make sure that happens but kind of going back to the national security thing is if we can stay ahead of this and we can lead it which i think we are the most innovative country in the world um we can do it and and that's what's really important across now the energy industry and the data center industry if you think about it those two things are becoming one yes um and you know these hyperscalers are now you know living in the industrial heavy industrial world and the heavy industrial world for a long time has been one very sophisticated on energy both power and gas and two a lot of them have operated on-site generation and very actively interacted with the grid and doing it and that is going to happen there's no choice the amount of power that's available [01:23:13] Kirk: on the grid is you know there's not a whole lot of that there so i think i appreciate that and thanks for bringing that out because i i want people to come off this podcast not only no longer worried about energy but hopefully inspired about the opportunities that energy brings huge yeah this is not a time to stop what we're doing this is a time to write policy around what we're doing and govern through what we're doing and that's the same pattern that we saw in the previous three transformational infrastructure revolutions we've ever seen and i just to remind people we when we rolled out the railway system we were worried about energy and demand and and impact to the environment from putting those tracks all over the world um we saw the same when we electrified the grid you know energy how are we going to generate it now how are we going to transmit it um those are very valid concerns and they still are today those challenges and those concerns haven't gone away um but the impact to the environment remain the top two things energy and impacting environment third was the automobile and it was more what is the impact of energy demand because of this and we know that the response was the consumption of gas increased by 50 fold in the first 10 years yeah but so did we add 3.7 million jobs and 80 of those came from small businesses so we created entire ecosystems of industry that were driving innovation and now we're using the innovation of technology itself the level the playing field [01:24:40] Dan Crosby: between small businesses and large ones i think for me like one of the most exciting parts of what's going on here you know i think we're kind of dispelling the myth that anybody in energy doesn't love the environment i'm a huge outdoor guy i grew up mountain tree rafting whole nine yards love the environment i don't think anybody in the data center space where the energy industry is out there to make a mess no one wants to make a mess but i think what's also so different about this revolution if you look at you know the boom of the internet you needed to be a coder you know and you needed to have you know really go to college and figure it all out or a lot of people didn't go to college and and figure it out that way if you look at the breadth of what's available to people to get into an industry that is you know very very very well funded maybe more well funded than anything we've seen certainly in a generation maybe in our lifetime is you can get involved from all the way down here at the trades which oh by the way are making a fortune yep um as you know high school education let's go you know and whatever the training program is all the way through being i mean if you want to go into engineering if you want to go into illegal if you want to go into environmental if you want to go into governmental policy if you want to go into utility and government regulatory the whole spectrum is available with high paying quality jobs and something that's also bringing massive tax base right and so if you think about you know there's all this talk about circular you know deals and everything that's going on and i'm not a finance guy per se but you know there's a lot rising the tide here of a lot of people that are gaining spending power that wouldn't have been able to do it otherwise and literally anybody from any walk of life can get involved at whatever point of that spectrum that [01:26:25] Kirk: they want to and that's exciting you know yeah no i agree i well just to bring this one home as we landed you know just to slightly touch on that last part that you mentioned with regards to the fears that were driven in this narrative regarding circular financing think of it like this ai is the most advanced weapon machine and tech ever built it's not built for the consumer first it's going to be built for national security which means it's too big to fail so when you see concerns and this guy's writing about how the you know the mouth is biting its own tail because we're just passing po's between nvidia and oracle and and open ai and a dozen other groups the reality is is the government's going to have no choice but to step in and govern this industry as it emerges into the mainstream will be more regulated than ever before and in many ways it's going to be safer it'll drive greater efficiencies just like we saw in the automobile industry just like we saw in the uh automation or uh airplane industry yeah where planes are safer to get into now than ever before because we drove safety requirements and efficiency requirements throughout that industry and the industry responded to it and the market did as well in this situation it's going to take a lot of help and support leaning in from the government and um i think that this industry is too big to fail now because of a national security perspective and i think that by the time that we're done explaining to you know our seat representatives that if we don't protect and defend this industry then we outsource these jobs and this technology overseas to places that don't share our values with regards to environmental impact or even what is best for the world in terms of peace and safety so we have to stay ahead of this we have no choice but to build through it we've proven as a nation it's the american way we did it with the railway lines we did it with the electrification of our us grid and we continue to do so with that and we do it with automation i mean with the automobile industry today as we continuously are are investing non-stop into infrastructure like better roads and better modes of transportation safer modes of transportation and more efficient modes of transportation so these things these same concerns are just the pattern of what happens during transformational infrastructural revolutions when it's going to touch everything [01:28:36] Dan Crosby: i mean and you think about how many really dangerous hard jobs a lot of people are doing right now they could be replaced by robots and those people can now get involved in this industry right [01:28:46] Kirk: and now safety they could go build those robots right exactly or they can do the maintenance on those robots but like i said small businesses will be the greatest benefactors and beneficiaries of this ai boom and we're already seeing it now so without these small businesses there is no ai boom so we need to get this collective industry almost unionized we need an international brotherhood of electrical workers for our industry we need this as a trade now we need data center apprentices and we do data center journeymen to go help deliver this product because if we don't do it then it's going to be outsourced and that's going to impact national security yeah dan thank you for spending time helping challenge these misconceptions people should now understand like it's your responsibility to understand what our industry is doing and how it's actually not consuming energy from the grid it's partnering with the grid to put energy back on the grid and we're going to re this industry will be responsible for upgrading the grid infrastructure that's going to make this industry more reliable which is going to make the communities that we live in more reliable from our perspective so thank you for taking the time to join the revolution pleasure man amazing time i do believe that the the data the ai factory revolution is mirroring 100 years ago to what the automobile revolution gave us as a as a country and i think right now it's our chance to build back the middle class for these jobs that are being demanded by energy so thank you for making the time and thanks for joining the revolution dan always pleasure man thanks

Transcribe Any Video or Podcast — Free

Paste a URL and get a full AI-powered transcript in minutes. Try ScribeHawk →