Data Center Debate: Are Energy Bills About To Explode?

Ed Elson speaks with Jigar Shah and Jon Parrella about whether the AI data-center buildout can really happen at the scale Big Tech and investors expect. They discuss why power is becoming the biggest bottleneck for AI, and propose solutions to ensure that America can build and manage the necessary power infrastructure. Jigar Shah is the former director of the Loan Programs Office at the U. S. Department of Energy. Jon Parrella is the CEO of Terraflow Energy. Get your tickets to the Prof G Markets tour Subscribe to the Prof G Markets Youtube Channel Check out our latest Prof G Markets newsletter Follow Prof G Markets on Instagram Follow Ed on Instagram, X and Substack Follow Scott on Instagram Send us your questions or comments by emailing [email protected] Learn more about your ad choices. Visit podcastchoices.com/adchoices

Speaker names marked ? are inferred from speaking time and the show’s known hosts — they are not voice-verified.

• 0:00Speaker 00

  Are your ad campaigns lighting up the dashboard? But not the pipeline? That's bull spend. And marketers are calling it out in dashboard confessions.

• 0:09Speaker 01

  My boss asked for results, so I opened my dashboard for the only positive sounding metric I had: impressions.

• 0:17Speaker 00

  Cut the bull spend. See revenue, not just reach. LinkedIn delivers the highest return on ad spend of major ad networks. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit. Go to LinkedIn.com/campaign. Terms and conditions apply.

• 0:30Speaker 02

  Energía baja. Falta de concentración? Nebla mental? You might be dehydrated. Gatorade Zero Powders han sido diseñados científicamente para ayudarte a mejorar el balance de hidratación en el cuerpo. Hechos con la mezcla número uno de electrolitos, sin colores ni sabores artificiales. Shop now online or at retailers nationwide. Gatorade: Hydrates better than water. Is it in you? Signos de dehidratación leve y moderada. Consulta a tu médico si los síntomas persisten.

• 1:00Speaker 03

  Welcome to Sephora.

• 1:01Speaker 04

  I'm looking for a perfume that's not too perfumey?

• 1:05Speaker 03

  I got you.

• 1:06Speaker 05

  Serum moisturizer or moisturizer serum?

• 1:09Speaker 03

  Let's get into layering.

• 1:11Speaker 04

  My concealer is making me look worse?

• 1:14Speaker 03

  Sounds like the wrong shade. Let's get you matched.

• 1:18Speaker 06

  There's only one store that really gets what you're going for. Get beauty from people who get beauty. Only at Sephora.

• 1:25Speaker 07

  Hi, I, uh...

• 1:26Speaker 03

  Let's get you a basket.

• 1:29Ed Elson?

  Today's number 400. That's how many meters long the world's largest ever tiramisu was, a record set last week by a consortium of 100 Italian chefs. Italy still faces severe economic challenges, such as crippling debt and a stagnant economy, but points for sticking with the whole food thing.

• 1:52Speaker 09

  Money markets, Matt. If money is evil, then that building is hell.

• 2:02Ed Elson?

  Welcome to Prof G Markets. I'm Ed Elson. Apologies to our Italian listeners. It is April 29th. Let's check in on yesterday's market vitals. The S and P and NASDAQ both fell due to a Wall Street Journal report that OpenAI missed its revenue and user growth targets. That news sparked a tech sell off with Nvidia falling more than 1% and Oracle and CoreWeave dropping 4%. Notably, Microsoft actually rose 1% ahead of its earnings, which are due tonight. Meanwhile, oil prices rose as negotiations with Iran continued to stall and the Strait of Hormuz remained shut. And finally, the United Arab Emirates announced it is leaving OPEC,

• 2:44Ed Elson?

  putting more pressure on the oil cartel's already strained supply.

• 2:50Ed Elson?

  Okay. What else is happening? Alphabet, Amazon, Meta and Microsoft are expected to spend more than $650 billion this year on data centers. And by 2030, data centers are expected to use twice as much electricity as they do today, enough to power France and Germany combined. The problem, though, is that the grid might not be able to handle it. Nearly 2,300 gigawatts of generation and storage capacity are currently stuck in the pipeline, more than the country's entire installed power capacity. Meanwhile, Americans are already seeing their power bills climb. In fact, residential prices could rise 15 to 40% over the next five years.

• 3:35Ed Elson?

  As a result, 14 states are considering moratoriums on data centers. So this leaves us with a multi-trillion dollar question. And that is, can the AI build out actually happen the way Wall Street hoped that it would? Well, to discuss this, we are joined by a panel of experts. Jigar Shah, former director of the Loan Programs Office at the US Department of Energy, and also John Perella, CEO at TerraFlow Energy. Jigar and John, thank you so much for joining me. Jigar, I'll start with you. When you look at the AI build out right now and sort of the obstacles that are in its way,

• 4:16Ed Elson?

  between compute, between people's feelings and political sentiment towards the technology, and then also energy, what do you consider to be the most important obstacle? The thing that's most in its way?

• 4:32Speaker 10

  Yeah, you know, it's such a crazy thing, right? You know, when you think about the five different bottlenecks you have to data centers, right? So there's the grid, right? Then there's transformers, then there's GPUs, then there's memory, and then there's CPUs now. It looks like they're short, right?

• 4:49Ed Elson?

  Right.

• 4:50Speaker 10

  They can't build more than 50 gigawatts of data centers between now and 2030 because of limitations on GPUs, memory, and CPUs, right? But if you go to communities, they're creating havoc at the level of 500 gigawatts across the country. So they are making empty promises to people across the country to disrupt their communities when they don't have the GPUs to fill it, right? And so that's, I think, what's pissing everybody off. I think people are just saying, if you're going to be a trillion dollars in size and you're going to be buoying the entire US economy, how come not a single analyst actually seems to know anything about what's possible?

• 5:32Speaker 10

  And everyone else is basically like feeding the hype cycle all day, right? I'll give you an example. In the state of Texas, they are suggesting, unofficial ERCOT letterhead, which is the transmission operator in Texas, that the load queue could be 300 gigawatts, right? Now, mind you, Texas is 70 gigawatts. That's a BS number.

• 5:56Ed Elson?

  Right.

• 5:57Speaker 10

  Right? And of course, people in Texas are scared out of their minds around a BS number. But what's likely to happen in Texas? 30 gigawatts.

• 6:06Ed Elson?

  Right. Yeah, John, I mean, this is coinciding, this conversation with some pretty significant news about OpenAI, which is that OpenAI missed its revenue targets last year reportedly, same with its user targets. And it reminds me of a lot of the promises that they have also been making. My understanding is that they were going to build 30 gigawatts worth of compute by the year 2030. I think they've built, I want to say, 2 gigawatts so far. I mean, the point being, it seems like that company specifically is probably guilty, in particular, of this overpromise, underdelivered dynamic that Jigger is talking about. I wonder if you have any thoughts on that?

• 6:49Speaker 10

  So I don't know that I quite agree with Jigger on 30 gigawatts. I think it's going to be a little bit more than that. We're working on quite a few of those projects in Texas. But

• 6:58Speaker 11

  And I don't know that I would put all the blame necessarily on OpenAI. Right, I I used to work for Lancome, that was the the the base land and power for the Stargate One project. I did a lot of work on that project before we knew it was going to be Stargate One.

• 7:12Speaker 10

  Yeah.

• 7:13Speaker 11

  And in in going back and looking at it, they've had a lot of issues because a lot of these companies, when they applied for their their grid interconnect, they applied as if they were like Bitcoin mine or traditional data centers that were very flat, stable loads. Um, if you actually start to look at the load profile of an AI data center, it's one of the most volatile loads you've ever seen. Um, 12, 14 times a minute load swings that are 30, 80% the size of the data center. And that's wreaked havoc on these data centers moving quickly because it's tearing up gen sets. They're literally breaking the crankshaft of the gen sets. They're burning through batteries. Uh, and it's it's a lot of these data centers that are all about time to power and moving quickly and they got ahead of the tips of their skis because they didn't architect the infrastructure correctly

• 7:58Speaker 11

  for it to succeed. And so I know that that even the Stargate project has had a lot of problems where the utility turned them back off. And they've, you know, had to scramble to try and find equipment uh, to be able to try and solve for these problems and and get them stable loads. So I I don't know that I would necessarily blame everything. It was definitely ambitious uh, on their part to be able to say that they were going to scale that big, that fast.

• 8:24Speaker 11

  I think in a lot of ways, they had to do that in order to get capital to be able to start scaling. Uh, but I think that they've also been hit by a lot of roadblocks where interconnection takes a lot longer than anticipated because these loads are big and complex. And I think that when they actually started turning these things on, what they expected and what actually happened were different. And that they started breaking equipment and finding replacements for the broken equipment is not easy right now. So it it just, it it's a, it's a whole cluster, if you will, of problems. So I while I I do believe that they were a little over ambitious, I do not believe that I would blame them 100%.

• 8:57Speaker 10

  It seems like a four-letter word that's after the word cluster.

• 9:01Speaker 11

  Yeah, well, I didn't want to quite go there, but

• 9:05Ed Elson?

  Someone can do it. Someone can break the ice.

• 9:10Ed Elson?

  Well, John, I guess.

• 9:11Speaker 11

  Trying to be a little bit more family friendly.

• 9:14Ed Elson?

  Clutter fuck, I said it. Okay.

• 9:16Speaker 11

  Yeah, there you go. All right.

• 9:18Ed Elson?

  So I guess, John, I mean, it sounds like what you and Jigger both agree that this is a lot more difficult than AI companies and investors had anticipated. It sounds like maybe where you may disagree is the extent to which they underestimated things, or the extent to which they misjudged it. I guess my follow up question to you, John, is, I mean, how bad is it really? From when we look at what's happening from Wall Street's perspective, for example, I don't think Wall Street is as clued into what is happening on the ground when it comes to energy and compute capacity constraints as you are.

• 10:01Ed Elson?

  So, I mean, how big of a problem is it really? If it's not quite as big as Jigger might be suggesting?

• 10:07Speaker 11

  So, I mean, I definitely think that there's, power's the number one issue. There's no question about it.

• 10:13Ed Elson?

  Okay.

• 10:13Speaker 11

  The thing that everybody's talking about is how long it takes to interconnect. Nobody's talking about what happens when you actually turn them on. And I think that obviously that as you see, you know, buying generators, there's a long lead time on getting gen sets. And you're seeing some deals come and some deals go. What you're seeing, it was no different before data centers were here. It was the battery craze, right? Like interconnection queue for people that went and had land and they could get power on land to interconnect and put a battery in. You know, it was like half the projects would never see the light of day, but everybody was connecting because if they could get the power to it, they could then flip it to a developer that would build it. I think you're seeing a very similar,

• 10:54Speaker 11

  at least we've seen this, where lots of people that have, you know, rural lands and there's all of a sudden a big power line that comes through their land, go, I can build a data center here. And, and, you know, they might have been able to get $5,000 or $10,000 an acre, but if they can get a large load interconnect, they can get $100, $200, $300,000 an acre for the same land. So you see a lot of people that came in and said, okay, if I can find the sweet spot of land where I can get gas and water and power and, you know, all these things, all of a sudden the value of that land goes through the roof and the data centers are willing to pay for it.

• 11:30Speaker 11

  What's funny, though, is when you start looking at what they submitted as their load profile for those interconnects versus what the data centers are actually doing. And so the problem is now the utilities are trying to do these feasibility studies off of, you know, a legacy, flat, very stable data center. And then these things show up and it's far from that. And that's, I mean, to be a little bit selfish, that's what our battery technology solves for.

• 11:57Ed Elson?

  Yeah.

• 11:58Speaker 11

  But it is, it is a major problem. And, and the reality is, you know, even if I stand up gigawatts of manufacturing capacity of batteries, I can't make enough.

• 12:06Ed Elson?

  Yeah.

• 12:07Speaker 11

  And that's the problem across the board with all of the infrastructure. And back when the Bitcoin mine days were here, and it was all the craze of building these monster Bitcoin mining data centers, you saw companies investing in, in, you know, PDU manufacturers and all the components because they wanted to prioritize their company over everybody else to be able to get the components. Same thing's happening with the data centers. The infrastructure funds, if you look at the Blackstones, are investing in TDSPs or utilities. They're investing in, you know, the, the, the land developers, they're investing in the infrastructure components, they're investing in all the different components to be able to try and prop up and accelerate the growth of all this to meet that demand.

• 12:43Ed Elson?

  So Jiger, I mean, when I kind of synthesize what's happening here, what we're, what we're basically saying is these data centers are a lot more volatile than we had originally expected. They're also consuming huge amounts of energy to a degree that people a few years ago would never have anticipated. And at the same time, I also know that the Strait of Hormuz is closed and blockaded as we speak. I know that gas prices have risen 30% over the past two months since we launched the war in Iran. I mean, all of these things together make me think we're basically headed towards a full-blown energy crisis.

• 13:25Ed Elson?

  Am I getting this wrong? What do you make of the picture, the energy picture in America right now?

• 13:30Speaker 10

  No, you're not wrong. I mean, you know, the data centers have a approval rating that's lower than the Iran war amongst the American, you know, people, right? And, and all of it is, comes down to basic selfishness, right? Like, when you look at our grid, if the data centers were willing to be flexible with their load for 100 hours, you could accommodate 100 gigawatts of data center load growth in our grid easily and reduce everyone's electricity bills by 10%. But that would require a coordination between the data center and the utility and the governors. And, you know, folks would have to like each other. But instead, they're saying, screw you, we'd like to, like, be run off-grid, right?

• 14:14Speaker 10

  Okay, that feels very selfish. I think when you look at where we are with the Straits of Hormuz, the US is in an extraordinary spot, right? We have a lot of oil production, we have a lot of refining capacity, that kind of stuff. When President Trump trumpets how many tankers are, like, heading towards Texas to pick up product, they're picking up product to take it to other continents, right? So they're picking up jet fuel, they're picking up diesel, they're picking up gasoline. And, you know, because we got a lot of it. Now, if we start shipping it all to other places, our prices are going to go up.

• 14:46Ed Elson?

  Right.

• 14:47Speaker 10

  Right? And you already see that. Diesel's up to like, you know, $6, $7 a gallon, gasoline is above $4 a gallon, right? That means that for a minivan, it's $85 to fill up that minivan. That's a lot, right? And so, when you think about 4.5 million households in the Northeast are still using fuel oil to heat their homes, that fuel oil comes from the Straits of Hormuz. And so now, they don't have fuel oil, right? And so, there's going to be a disruption around the world, we're going to have a humanitarian crisis around the world. The US is better situated than almost everybody else is. But, make no mistake, right? We're not putting fertilizer down for a lot of our land, right? Or we're going to put only half as much down. That means yields will go down.

• 15:31Speaker 10

  That means food prices are going up, right? And so, part of the challenge I see right now is there's a way to use the best technology in the world, invented here in the United States, deploy it at scale, to give people relief. But, when you have a president that's panning people off to shut down offshore wind farms... Right? You're not using the best technology in the world. You're instead figuring out how to like make the crisis worse. Right? And so we have got to scale up solar, scale up battery storage, scale up electric vehicles, scale up all of the things that takes the pressure off of fossil fuels.

• 16:10Ed Elson?

  Stay tuned for more of this panel after the break. And by the way, we are heading out on tour at the end of May. So for more info and to get tickets to a show near you, head to profgmarketstour.com.

• 16:29Speaker 00

  Are your ad campaigns lighting up the dashboard? But not the pipeline? That's bull spend and marketers are calling it out in. Dashboard Confessions.

• 16:38Speaker 01

  My boss asked for results, so I opened my dashboard for the only positive sounding metric I had: impressions.

• 16:46Speaker 00

  Cut the bull spend. See revenue, not just reach. LinkedIn delivers the highest return on ad spend of major ad networks. Advertise on LinkedIn. Spend $250 on your first campaign and get a $250 credit. Go to LinkedIn.com/campaign. Terms and conditions apply.

• 16:59Speaker 02

  Energía baja. Falta de concentración? Nevra mental? You might be dehydrated. Gatorade Zero Powders han sido diseñados científicamente para ayudarte a mejorar el balance de hidratación en el cuerpo. Hechos con la mezcla número uno de electrolitos, sin colores ni sabores artificiales. Shop now online or at retailers nationwide. Gatorade: Hydrates better than water. Is it in you? Signos de dehidratación leve y moderada. Consulta a tu médico si los síndomas persisten.

• 17:29Speaker 03

  Welcome to Sephora.

• 17:30Speaker 04

  I'm looking for a perfume that's not too perfumey.

• 17:34Speaker 03

  I got you.

• 17:35Speaker 05

  Serum moisturizer or moisturizer syrup?

• 17:39Speaker 03

  Let's get into layering.

• 17:40Speaker 04

  My concealer is making me look worse?

• 17:43Speaker 03

  Sounds like the wrong shade. Let's get you matched.

• 17:47Speaker 06

  There's only one store that really gets what you're going for. Get beauty from people who get beauty, only at Sephora.

• 17:54Speaker 07

  Hi, I...

• 17:55Speaker 03

  Let's get you a basket.

• 18:03Ed Elson?

  We're back with Profy Markets. I keep on hearing about how, you know, America is energy independent. We do it all on our own. And then it's like, well, why are prices rising as much as they are? And why are electricity prices rising as well? I mean, it's not just the gas bill, it's the electric bill too. And it makes me think, okay, well, I mean, maybe we're more independent than other nations, but it seems like we're not that independent. We're not so independent such that we are actually insulated from things that happen abroad. And we're certainly not independent enough that we're able to keep up with the demand of these data centers at the same time.

• 18:43Ed Elson?

  So I guess when you look at our energy picture, I mean, how independent are we really? And what would it take to get to a place where prices at least moderate or start to come down? Or where data centers don't cause a full-blown cost of living crisis among Americans today?

• 19:03Speaker 11

  So it's an interesting question, and I'll answer it in a number of different ways. One, I think if you look at the economy right now, data centers are absolutely starting to prop up the economy.

• 19:12Ed Elson?

  Yeah.

• 19:12Speaker 11

  If you look at the amount of jobs that they're creating, if you look at the amount of factories and made in America, I mean, my entire company right now is here and the jobs that I'm creating because of this data center boom that's going on. The cost of power is all about supply and demand. It's a real-time commodity. It's a real-time supply and demand. So there's no question that supply and demand and data centers, you know, when they first started, they were all about, I want grid power because it's very stable power. Right now, most of the data centers that were involved in, and we're working on over 100 gigawatts of projects across the United States and now going international.

• 19:53Speaker 11

  Right now, I would say probably 80% of them are planning to go behind the meter generation.

• 19:58Ed Elson?

  Could you explain what that means for our audience?

• 20:00Speaker 11

  So behind the meter generation means that instead of grid connecting, at least for phase one, they're going to bring in natural gas and put up generators or turbines on their campus, and they're going to generate their own power.

• 20:11Ed Elson?

  Yeah.

• 20:11Speaker 11

  And that was part of like that Trump pledge that he was really pushing is that you're going to build more generation. Now, think about this challenge for a second. All of the generation assets are spoken for, for the foreseeable future.

• 20:24Speaker 11

  You have companies like Microsoft and these big meta buying nukes right now so that they can power their data centers, which is the cheapest power out there, right? Coal, nuke, your baseload powers. And so they're they're co-locating their data centers with generation assets so that in some situations, there's grid constraint because the lines aren't big enough to carry it. But at the same time, they're buying the assets and almost removing them from the grid is is part of the builds, the bid stack of what determines what the price of power would be. Uh huh. Now, think forward a minute, and I know everybody's looking at it saying they're raising the price of power. Look, they're the SB six in Texas, Senate Bill six in Texas, forces the data centers to pay for all their own interconnect and all their own, you know, uh, equipment. So they're not passing that along to the rate payers everywhere.

• 21:09Speaker 11

  Well, and to be interruptible. But and and and listen, that that was where I was going. Now, if you look at architecting grids, and Jiger and I had this conversation literally two days ago. I personally believe that regulatory needs to come in and say, if you want to build big, you have to be a grid asset, and you have to be a responsive load rather than just a potentially controllable load. Let me explain the difference between the two. Yeah. Controllable is something where the grid can come and say, with 30 minutes notice, I need you to turn off for up to 12 hours. That's Senate Bill six in Texas. And what that means is that during emergency periods, they would make the data center shut down. Well, there are 13 times in the last five years that might have happened. Okay. A controllable load means that you can set up the grid to use the frequency of the grid

• 21:52Speaker 11

  to control the data center or or like a demand response program, if you will. So when the grid really needs it, the data center can control the load. Now, in many situations, for those that built behind the meter generation, if their electrical power infrastructure is set up correctly, it can enable them to actually inject power back into the grid from the generation assets that they have. So now, all of a sudden, all these data centers are building out all this power, and all the grids are building out all the infrastructure to connect them to the grid eventually.

• 22:19Speaker 11

  And now what you're going to do is you're going to see this huge revolution where there's more generation capacity available when needed that balances the grid. If you look at what Jigger said before about, you know, if you look at like 95% reliability from a utility, where you basically take the data centers off the grid during the five to seven peak hours of every day, in that situation, you could actually fit most, if not all, the data centers that want to be built on the grid today.

• 22:45Speaker 10

  And reduce everybody's rates by 10%.

• 22:48Speaker 11

  Right.

• 22:49Speaker 10

  Like, that's the thing that's so crazy. This, what John is suggesting is the unselfish way to participate in the system and achieve speed to power, right? Jensen Huang just said the same thing two weeks ago and said, "We have all the technology to do exactly what John is suggesting." And now the data center companies are mad at him because they're like, "That's not masculine enough for us. We'd rather run off-grid and be selfish." I was like, "What are you talking about? Like, we're giving you speed to power faster because the generators are sold out, and you'd reduce everybody's bills by 10% by using the grid we've already paid for more efficiently, right?

• 23:30Speaker 10

  But you'd rather be selfish than actually get connected faster and, you know, be a better neighbor. Of course, people hate you."

• 23:38Speaker 11

  Yeah. Yep. So it's really a path to how do you get to that, that turning these data centers into grid assets instead of grid liabilities. And that's a big part of what our product line does. That's a big part of what we've been trying to tout from the mountaintop, is there are ways to do it. Google's trying to lead the way by looking at all different things, everything from Emerald AI to be able to curtail load, to shifting load, to everything else. Now batteries are playing a huge role in enabling that. And that's, that's part of what we're doing, is building proper grid-scale energy storage that's long-duration, that enables the data centers to curtail load, and then if they need to wield the generation assets, they have the ability to wield the generation assets. It starts with them building behind the meter because the grid can't get there fast enough.

• 24:23Speaker 11

  But if they apply as a either controllable load, or responsive load, or demand-response programs, these, these huge liabilities become huge assets that will drive the price of power down. I don't agree that it's the data centers that are driving the price of power up. I think it's, the economy as a whole, okay? Um, yes, the data centers are contributing to it. No, I do not believe that everything that you see in the queue will be built. You have a lot of people that don't have the financial wherewithal that are just trying to, you know, ride on the gravy train. I think that the, the, the things that, that for instance ERCOT's enacting with deposits and everything else helps to weed out the fake ones and enable the ones that are real to get there faster. I think those things need to be done. And you've seen the president try to get FERC involved to accelerate because a lot of

• 25:08Speaker 11

  these grids are, are absolute cluster fucks. I'm, I'm gonna, I'll break it now. I'll say it. But like PJM can't get out of their own way. They can't get anything through the interconnection queue fast enough. Right? And, and that's, there's no reason for that, especially not with AI now. Like you should be able to accelerate the speed of feasibility studies leveraging AI's capacity to look at the network of the grid and look at it and say, is there room here or not? And how do we do it? The other thing I'll, I'll mention that that's interesting is if you look at what happened in Virginia recently, there's a misalignment or misunderstanding of certain key terms that, that the grids are reacting to. And, and it's scary because I think you're going to see a lot of like regulatory reaction and you're starting to see moratoriums because they're like, okay, we need to slow everything down a little bit and pump the brakes to make sure we fully understand what this means.

• 25:54Speaker 11

  In, in, in Virginia, they had a situation where it wasn't a fault of the data centers, but the grid faulted. And like nine data centers went offline or they went to backup power. And when the grid came back up, the grid was expecting the data centers to be there and they weren't, they were still running off backup power. It's called low voltage ride through. And, and what happened is the grid then had an over frequency event because the, the, it's all real time supply and demand. So you had generators online for the grid that were expecting a load and load wasn't there when the grid came back. And so there was way more power available than, than needed being what can happen in situations like that and/or if you connect one of these data centers directly to the grid, when you have that much volatility that quickly, you can shunt, uh, uh, uh, air breakers or, or trip air breakers at substations and cause a rolling.

• 26:38Speaker 10

  Well, you get rolling blackouts is what you get.

• 26:40Speaker 11

  Right. And so when you start building data centers that are gigawatt scale, it scares the hell out of me if they don't have the right infrastructure in place because they're not designed, they, they were moving too fast. They got above the tips of their skis and they built these data centers that are going to cause rolling blackouts. Then what do you think?

• 26:56Speaker 10

  Well, when, when the tech companies say they're moving fast and breaking things, normally they mean, you know, like little things. In this case, you're talking about people's lives, right? Like, when you think about how many people live on CPAP machines or other machines at home, they can't have rolling blackouts because a data center didn't plan correctly, right? And so, like, I actually think data centers are gonna reduce everyone's electricity bills by 10%, but they gotta be done the right way. And you need, like, a US Department of Energy that's forcing them to do it the right way, right? And so, one of my big challenges is that we're allowing people to move fast and break things, and, you know, like, this is not something that we can afford to break.

• 27:40Ed Elson?

  If I could just, I mean, a lot of technical terms here. I wanna just kind of, like, put it in layman's terms that even I can understand it. It sounds as though, it sounds as though the right way to do it, and let me know if I'm getting any of this wrong, is for the data centers to be connected to the overall grid and responsive to the demand that the grid observes, essentially. Excuse me if my language is wrong. But essentially, a data center that is in real-time responsive to the capacity and the supply and demand dynamics of the grid at large for a given geography, is that correct?

• 28:23Speaker 10

  It's basically two things, right? One is, is that, exactly as you suggested, right, that the grid has plenty of capacity almost all of the time, except for maybe 100 to 200 hours a year.

• 28:38Ed Elson?

  Yep.

• 28:38Speaker 10

  Right? During those 100 to 200 hours a year, the utility says to the data center, you gotta, you know, reduce your demand from the grid by 50%, right? And either they can run natural gas generators behind the meter to ride through that, or they can run it off batteries, or they could, you know, use Emerald AI and others and just shift the compute packages to a different data center, right? Separately, the data centers have to actually have good-quality load facing the grid, right? And we've set those standards in the 1960s and 70s, so this is not new. You can't flicker. You can't change how much power you're using from the grid 12 times a minute.

• 29:21Speaker 10

  Right, you can't. So then you have to have technologies like John's in between you and the grid, so that his technology is absorbing all of that flicker and the grid is only seeing a signal that it's used to seeing since the 1960s, right? And if you decide to send them something that's far more volatile than that, then the grid gets weird harmonics, it gets all sorts of things. And that's what John was saying before, that when the grid sees that, it tells the data center to shut down until it can run safely, right? And so, those are the two big things.

• 29:57Speaker 11

  Well, they're gonna trip the airbreakers is what they're gonna do.

• 29:59Speaker 10

  Well, that's right.

• 29:59Speaker 11

  That's, that's, like, what, what most people don't understand is when you have a power that's this volatile, where you're seeing, like, if you have a gigawatt data center and you see anywhere from a 30 to 80% swing, you're talking about 300 megawatts to 800 megawatts of swing multiple times a minute. What I mean swing, I mean immediately disappearing off for a couple of seconds, immediately spiking back on for a couple of seconds. It looks like an EKG up and down and up and down. And, and, and the problem is, think about this for a second. The fastest generators on the planet, you're probably talking diesel or recips, take 25 to 30 seconds to ramp up and ramp down. You start talking turbines, you're talking 10, 20, 30 minutes to ramp up and ramp down. Nukes are on all the time, they don't ramp. Solar is intermittent, it's on when it's on, it's off when it's

• 30:43Speaker 11

  off. Okay? So now, how do you take a, a very volatile load and connect it to those things? If you connect directly to a generator, like Jigger mentioned, where you're running off the generator behind the meter, and you don't have something between the generator and the data center, you're, you're literally gonna break the crankshafts on the, on the gen sets. It's like driving a, a supercar down the highway at 80 miles an hour and shifting from fifth gear to first gear over and over and over again. How many times is a car gonna let you do that before you drop your clutch, you drop your, you know, your, your engine just falls out and blows up, right? That's what they're doing to the gen sets. And it's it's literally breaking the grid, if you will, if they connect that. Now, the question you have to ask is, how many of these data centers you've heard about are actually online right now?

• 31:26Speaker 10

  Right. Zero.

• 31:27Speaker 11

  Very few, very few.

• 31:29Speaker 10

  There is not a single data center in the country that's running off grid. I just had Tim Hayde on my my podcast, Energy Empire, and like, yeah, zero, zero. Like, even Stargate is grid-connected right now. You know, they're hoping to be off-grid in the future a little bit, but right now they're fully on-grid.

• 31:49Ed Elson?

  Which is why it seems to me that while maybe, I mean, John mentioned that maybe the energy, the increase in energy prices might not be directly because of data centers right now, my assumption is when everything comes online, they certainly will be in the future. And in addition, if what you're kind of describing here is that these data centers aren't being really responsibly built, as in they're not setting up their systems such that they won't run into issues later down the line. And then I look at something like Allbirds deciding to rebrand themselves as data center providers, or even, I think, Jigga, you mentioned the name Fermi there, which is a company that we were looking at, which seems to be a total con, it seems to me.

• 32:33Ed Elson?

  And then we saw that the CEO exited. I mean, I would like to hear your comments on those companies.

• 32:40Speaker 10

  Well, the other the other piece of this, just to make sure we're crystal clear, right, is that like, you've got the data, fine, but then you have the narrative, right? Every electric utility that's an investor-owned utility is publicly traded. When you listen to the conference calls of all of those CEOs the last three months, they are firmly blaming data centers for rate increases. They have added like 40% to their budgets since 2024, right? That means that they said before, we're raising rates 5%. Now they're raising rates 9%, right? And they're saying it's all because of load growth from data centers. Now, you could call them all liars, okay, but these are the electric utilities that

• 33:25Speaker 10

  actually serve you, right? They separately had a report last year that said they were going to spend a trillion dollars between 2025 and 2029 investing in the grid. This year, they just re-upped that from a trillion to 1.4 trillion from 2026 to 2030.

• 33:43Speaker 10

  Right? If they successfully do that, then that means higher bills for everybody. Now, if they allow the data centers to be more flexible for 100 hours a year, that number drops in half. It goes from 1.4 trillion to 700 billion, right? And you still get all that extra load from the data centers, which means people's bills go down by 10%. Like, that's the thing that people don't understand. This is just math. The numerator is the spending from the utility. The denominator are the kilowatt hours that you're selling.

• 34:15Ed Elson?

  Yes.

• 34:15Speaker 10

  You want that ratio to go into the direction of the consumers.

• 34:19Ed Elson?

  What is the correct policy response here? For when we think about the data center build out, and we think about the fact that they are becoming so unpopular, and that, I mean, everyone knows what a data center is at this point. They probably didn't two years ago. We're now seeing proposals for data center moratoriums. Just get rid of them altogether. And then on the other side, you have an administration which, in my view, has made a policy of having no policy, really. I mean, that there was an executive order to eliminate all forms of state-level AI legislation, which to me is basically them saying like, we don't want any regulation. We just want the markets to completely do their thing, no regulation whatsoever. To me, neither seems like the right move.

• 35:03Ed Elson?

  But I'd be interested to hear both of your perspectives on what is the right move from a policy perspective. Jigar, I'll start with you, considering you recently worked in government.

• 35:12Speaker 10

  So there's two big things you have to do. And, you know, Texas has passed SB6, as John suggested, which basically says that if there is ever a shortage of power in the state of Texas, data centers go offline first, right? And so that is, should be the law of the land, right? That's called an interruptible tariff. It's been around for 50 years. It is not new. And all data centers should be required to sign one, right? Second is that the utilities have a bunch of data trapped in their databases between three different silos. There are companies like Gridcare and Google Tapestry and Camu Energy and others who can use that data to tell you exactly where there's space on the grid today.

• 35:57Speaker 10

  And there's 100,000 megawatts, by the way, of space on the grid. So Gridcare just announced a deal with Portland General, where Portland General didn't know, but they had 480 megawatts of spare capacity in this one location. And I mean, those guys are ecstatic because they're unlocking 480 gigawatts or 480 megawatts, sorry, they're going to be able to lower everyone's bills by 10% in Portland General's territory, right? But because the left hand and the right hand don't talk to each other within the utility, they didn't know that they had that 480 megawatts there, right? That's it. Those are the two things. Governors need to say, everyone's on an interruptible tariff. I don't care how powerful you are and how much money you have to throw around. And everyone at the utility has to share their data with these AI-enabled software

• 36:42Speaker 10

  packages that can tell us exactly where you can put more megawatts on the grid and lower everybody's bills.

• 36:49Speaker 11

  John, does that fix it? Well, I think there's a couple of things. One, I would take it a little farther than SB 6, like I stated before, which is make them more responsive to be more grid participants. So more than just during emergency periods, but almost on a daily basis. Second thing is, when you look at the interconnection queue, again, if you look back to the battery craze, there's a lot of projects that have been approved for interconnection that are just sitting there because there are people with land trying to flip the deal. And I think that there needs to be more skin in the game, if you will, to a certain extent, to prove that you've got the financial wherewithal to actually move forward the project within a timeframe where you lose your interconnection, so that you can make room for other people

• 37:31Speaker 11

  that actually have projects that want to get done. I can't tell you how many substations. I mean, there's plenty of different applications, like Jigger was saying, where you can go, and we do this ourselves, where we can do feasibility studies to see what capacity is where. And especially with AI, you should be able to load that data in and get it to spit out and be like, there's megawatts everywhere. I think what you're going to see as far as data centers is going to shift. I think trying to get to gigawatt scale is a huge challenge and a huge lift to be able to get everybody on board and the regulatory and it's, they're very dangerous. We're seeing a lot of shifts very rapidly to edge AI. And a lot of people don't necessarily fully understand what that means. You're going to have AI corridors throughout the world, all right? And the AI corridors are going to be like major highways

• 38:16Speaker 11

  where you've got transmission lines and everything, where they build 10 or 20 or 50-megawatt little smaller data centers to enable things like automated driving or automated flying or things where you need low latency to be able to get to response time. And so to me, distributed AI, I think, is going to be more the way we're going to see things move quickly. There's already companies we're working with like Dell and these other big companies that are doing all these like, containerized AI in a box, just like Bitcoin got to, right? Where you could find pockets of power and you could deploy smaller data centers to get the GPUs working faster. And a lot of those get underneath the hurdle of the regulatory where you're, you know, Texas, it's 75 megawatts and above is large

• 39:01Speaker 11

  load. So if you build, you know, 5, 10, 25, 50 megawatt smaller data centers, you can build them much faster, you can get them online faster, and you create less of a liability to the grid. And that's where you find pockets where you can hyper-locate next to substations and use up the capacity of those. I think that's the way we're going to see.

• 39:20Speaker 10

  Yeah, you saw a big announcement between Strata Expanse and available infrastructure and Data Vault for a $5 billion rollout of distributed AI. So it's happening now.

• 39:31Speaker 11

  And so we see not, because here's the other problem. If you look at the Stargate One project, by the time they completed building the first few buildings, they were already like four GPU generations behind. Right? And so the difficulty with building these massive structures and building a whole city around it, is one, the staffing you need in a specific location, then two, the capital, and then three, the lead time to everything. When you do modularize like AI in a box that you can just drop the container off, and when that container becomes out of date, you pick it up, take it back to the factory and drop another one off, I think you're going to see kind of the AI factory change. And listen, I just got back from Data Center World Conference in Washington,

• 40:15Speaker 11

  DC, and we saw just that. There were, like, where there was maybe one, now there's like seven or eight containerized AI companies now building that. So we're seeing a lot of that as well.

• 40:25Ed Elson?

  Yes. Many data centers, which may hopefully make us less reliant on a handful of companies to build all the data centers that we need. I mean, it seems as though, literally, four companies are going to build the giant robot boxes, they're going to power our AI. What if you could build smaller robots that other people and other companies could own and profit from? Jigar Shah, former director of the Learn Programs Office at the US Department of Energy, John Perella, CEO at TerraFlow Energy, John Jigar. I appreciate your time. Thank you.

• 40:56Speaker 11

  Thank you very much.

• 40:57Speaker 10

  Thanks, Ed.

• 41:01Ed Elson?

  Okay, that's it for today. We appreciate you joining us for another Prof G Markets panel. If you have a guest you think we should speak to on this topic or any other, please drop us a line in the comments or email our producer Claire at [email protected]. We hope to hear from you.

• 41:19Ed Elson?

  Okay, that's it for today. This episode was produced by Claire Miller and Alison Weiss, edited by Joel Patterson and engineered by Benjamin Spencer. Our video editor is Brad Williams, our research team is Dan Shalane, Isabella Kinsel, Kristin O'Donoghue and Mia Silverio, and our social producer is Jake MacPherson. Thank you for listening to Prof G Markets from Prof G Media. If you liked what you heard, give us a follow. I'm Ed Elson. I will see you tomorrow.