Volts published eight conversations in August — with two Idaho state legislators, a grid cybersecurity expert, the architect of Uruguay's energy transition, an EV charging exec, and more — which is a lot for anyone to listen to. We've rounded up the highlights into a single hour. Enjoy!
So last month, we tried something new around here, a compilation of clips from the month’s episodes strung together in a single place so that listeners could scan and see what they might have missed. I asked at the end of that one whether you found it worthwhile and useful, and many of you wrote in to say that you did, so we’re going to make it a regular feature.
Here’s how it will work. Every month there’s a best of episode. Paid subscribers get early access, and when the next one comes out, the previous one opens up to everyone. The July edition is open now, so if you haven’t heard it or you know somebody who’s been meaning to check out the show, that’s the one to send them.
As for August, it was wide-ranging, which is how I like it.
Data centers twice from opposite ends, including state lawmakers dealing with them and companies trying to improve them. Grid cybersecurity, public EV charging, a surprising run of bipartisan housing bills out of Idaho, why utilities can’t seem to adopt anything new, what a future administration ought to do with the tools this one is busy beating us over the head with, and a listener favorite out of Uruguay.
As always, I deeply appreciate all of you who support this work by becoming a paid subscriber. This podcast is 100 percent listener supported. I could not do it without you.
Chapters
00:00 – Introduction and how the best-of series works
01:59 – Saleem Chapman on why every market design fails ratepayers the same way
07:51 – Patrick Miller on Chinese inverters and commodity boards
12:22 – Ben Toews and Megan Egbert on how Idaho's housing bills happened
18:34 – Seth Cutler on why an IONNA site is technically a bakery
22:52 – Ramón Méndez Galain on getting every Uruguayan party to sign one plan
28:00 – Quinn Nakayama on why a utility would adopt something that saves money
34:16 – Varun Sivaram on workload tiers and who pays when flexing goes wrong
38:44 – Todd Tucker on the Defense Production Act precedent
44:20 – Closing note and what to send in
Resources
People & Organizations
David Roberts (Substack - Bluesky)
Volts (Apple Podcasts - Spotify - YouTube)
Saleem Chapman
Patrick Miller (LinkedIn)
Ampyx Cyber (Website)
Ben Toews (Idaho Legislature)
Megan Egbert (Idaho Legislature)
Seth Cutler
IONNA (Website)
Ramón Méndez Galain (Climate Breakthrough)
Quinn Nakayama (LinkedIn)
PG&E Research & Development (Website)
Varun Sivaram (LinkedIn)
Emerald AI (Website)
Todd N. Tucker (LinkedIn - Roosevelt Institute)
Company & Industry News
Books & Articles Discussed
Patrick C. Miller - Written testimony before the US-China Economic and Security Review Commission
Roosevelt Institute - Building Up in 2029: How to Make Green Statecraft Durable
Roosevelt Institute - Trump Wields the Defense Production Act to Promote Fossil Fuels
Related Volts Episodes
Grid cybersecurity: how big is the threat & what should we do about it?
He helped Uruguay decarbonize. Can he help other countries do the same?
Making data centers flexible so they can serve the grid rather than stress it out
Text Transcript
David Roberts: All right then, with no further ado, we begin with Saleem Chapman of Climate Cabinet on why a variety of electricity markets are failing on data centers and how states could do better.
One point that I think is important in the paper, and that is interesting to me, is that these risks, they’re not necessarily unique to a particular kind of market design. You know, as Volts listeners know, there are vertically integrated utilities some places. There’s what are called restructured markets. There’s ERCOT, whatever ERCOT is. One of the points you make in the paper is that people are making kind of mistakes. There’s this kind of the same failure pattern shows up in all kinds of market designs. Can you walk through that a little bit? Maybe some examples of, like, the different places that are, that are messing this up.
Saleem Chapman: Sure, yeah. So we, in this part of our paper, we looked at three different markets, that operate within three different states. And so if you start with, Georgia, which is the example of a regulated energy market-
David Roberts: Right. Old school, vertically, old school, vertically integrated ...
Saleem Chapman: Old school, vertically integrated, I think those of which most people would sort of be most familiar with and think about when they think about, utility structure. And so what we saw in Georgia is that the, Public Utility Commission approved a major gas expansion really built on justification that was derived from developer supply forecasts. And so in that instance, the process worked exactly as it should. There was nothing sort of bad faith about it, but it still, created a situation where ratepayers had significantly exposure, cost exposure, if those demand projections ultimately fell short.
David Roberts: Right. So the ratepayers are paying for the gas, the gas plant build-out, and if the demand doesn’t show up to absorb all that gas build-out, it’s the ratepayers who are paying for the stranded assets, basically.
Saleem Chapman: Exactly.
David Roberts: And then PJM though, PJM is a, is a competitive market, restructured market. Talk through that one a little bit.
Saleem Chapman: Yeah, so what we are seeing in a state like PJM is that you can have a single hyperscale, campus being brought online. At the cases that we looked at, it was 30% of local peak, demand within that particular region. And that triggered a capacity market event which ultimately resulted in, as you’ve-- if people have been following the PJM sort of issue, that you’re seeing sort of historical capacity prices year over year. And so what we are seeing there is that similar, the process worked exactly as it should, but by triggering that capacity market event, you had states all across the 13-state footprint that PJM covers having to take on some of the cost, that was associated with bringing that new facility online when they had no say in actual the decision-making around whether or not that facility would be approved.
David Roberts: Mm-hmm. Yeah. Who pays? This comes up again and again. And then there’s finally ERCOT, which people, you know, I think which people think of intuitively as, like, sort of the most free market, closest thing to a free market type of system. And they might think, well, in that kind of market, surely, you know, it’s the developer who takes the risk if it’s, you know, really working like a free market. But it’s, turns out there’s failures there too. Talk about that one real quick.
Saleem Chapman: Yeah, exactly. I mean, what we also know is as we are confronting this new period of technological innovation, we are also operating within a much more volatile climate than what we have historically had. And so in the ERCOT case, we saw hyperscale growth, colliding with extreme weather events. And so some of the emergency programs and gas dispatching that had to be put in place to meet that emergency event added billions in surplus charges that really surfaced as price volatility, but wasn’t actually visible as a line item on any individual ratepayer’s bills that said, you know, “This was driven by hyperscale growth.”
David Roberts: Right. Right. It shows up as price volatility in the market, which is absorbed by everybody who’s involved in the market.
Saleem Chapman: Exactly.
David Roberts: And so all these kinda come down to who is taking the risk and who is paying the price. So in all these cases, you take data centers’ demand forecasts, these developer demand forecasts, translate them into a bunch of costs, right, to anticipate meeting that need, and it is the ratepayers ultimately who are paying those costs to prepare. And so in other words, it’s the ratepayers who are taking on that risk. It’s the ratepayers who are paying. And then if the developer forecasts don’t pan out, the ratepayers are stuck with it. That’s like, that’s kind of the core theme through all those market, despite the market structures.
Saleem Chapman: That is the core theme, that ultimately this comes down to who’s taking on the risk and who’s capturing the value.
David Roberts: Next up, we have Patrick Miller, cybersecurity expert on the Chinese inverters that everyone is worried about.
As I think people probably know, most inverters, you know, we’re talking about inverter-based resources. Most inverters come from China right now. There was this big Reuters investigation in 2025 which said that inspectors found all kinds of little undocumented communication devices inside solar inverters and batteries, and I guess, like, t-two questions. A, what do you do about what I imagine are now hundreds of thousands, if not millions, of those Chinese inverters that are already installed somewhere and running? Like, you can’t yank them all out. You c- there’s no practical way to audit them all. What do you do about those? And then what do you do about the new ones? You c- as you say, we have no jurisdiction over China. Like, how do we exercise any control over what China puts in those inverters?
Patrick Miller: Yeah. I’m gonna keep this at a non-technical level ‘cause it gets really technical really fast. So in general, you’re absolutely right. There’s a, there’s millions of these things out there, and they do have a lot of undocumented communications components in them. But I wanna make sure we’re really understanding what that means without any hype whatsoever. When China manufacturers a lot of this equipment, they literally do it on what’s called a commodity board. So you can imagine in your mind like a motherboard, right? It’s a- you know, it’s a, it’s a computer component where the chips and the memory and all those things sit, and they’ll buy literally millions of these at once. Mm-hmm. And they’ll buy them, you know, from a commodity manufacturer that just makes, you know, boards, literally, like generic board. And on that board, it’ll have all the things just in case you need them, whether you use them or not. They don’t buy specific boards with just these components. So when they g- uh, I mean, 99% of the ones I’ve seen that they’ve gone out and found stuff, it’s just been the commodity board. Yes, it came with a radio, it came with a modem, it came with a place for a SIM card. There wasn’t a SIM card inserted, but it has all those components on there in case the buyer of these commodity boards wants to use that functionality.
David Roberts: I see. I see. So not nefarious inclusion of these, right?
Patrick Miller: Not necessarily. So whenever you see those, you kinda have to take that with a grain of salt and think, well if there’s millions of these things out there, did they just use commodity gear and they’re just not, you know this is just part of the board that maybe it’s used, maybe it’s not. Now, they have found some of these things phoning home and, you know, the routing traffic through China, so that’s a different discussion. That’s at the software level, you know, up where the applications that you’re using, to manage these devices and that kind of stuff. So those are, those are different components, and we, there’s some supply chain there that we can impose because it’s software, right? We can inspect code. We can enforce certain rules, on how code is developed and what’s, you know, certain, like, transparency of the code. We need to look at the code before we’re willing to allow it, these kinds of things. And we’re not just gonna let updates from China happen without someone taking a look at it before it goes in, those kind of things. Well, yeah, I mean, you- So there are some gates we can put in place...
David Roberts: Well, yeah, I mean, you say there might be these elements on the boards just because the boards are sort of generic and created for maybe lots of different purposes. Could those, could those components of the board that aren’t being used be activated later? Do you know what I mean? Sort of remotely activated. Like that’s, I guess, is the worry.
Patrick Miller: Yeah. They’re there, and they could be activated. But they’d have to have... I mean, you could activate them, but you’d also have to configure them. Like, it’s hard to explain, but when you, when, just because you turn something on doesn’t mean it’s ready to be used. Like even a phone, you can turn a phone on, but you have to give it a SIM card that’s got an activation code, that’s got, you know, all these things have to happen before it’ll actually work on a cell network. I mean, you could get a phone out of the box and turn it on and it just doesn’t do anything. So those pieces of gear are there. That doesn’t mean even if you did enable it, that it would automatically work. So it does, it ups the ante in terms of just like flipping a switch and getting this like rogue network of these things to go do stuff.
David Roberts: Then, on to Idaho, where legislators Ben Toews and Megan Egbert explain how each of them got radicalized on the subject of housing.
Nothing radicalizes people about housing in America more than trying to build something. Tell us how you sort of came to be engaged on this issue.
Ben Toews: Yeah, certainly. We recognized that there’s a serious issue with affordable housing, especially for the next generation to own their own home. The cost in our area is around $550,000 on average for a single family dwelling. And so, back in 2020, we purchased, a small piece of land. It had a PUD, so it had, 20 units already set up to be built on it, and, it was gonna be small homes, probably average 1,000 square feet. The range would be 700 to 1,200 square feet, and it was on small lots, and I was like, “We can sell these. We’re hoping to sell these in the $300,000 range,” open the door for the next generation to be able to get into a house. And, partway through the process, the city came back to us and we went to renew the PUD, and they said, “Sorry, you can’t do that. There’s new green space requirements, and your plans don’t, match those requirements.” And so they shut the whole project down. And so it was very frustrating. I wasn’t in the legislature at the time, but I saw it as a real issue back then, and I... and it just became more and more of an issue.
David Roberts: Awesome. And Megan, same question to you. You were a librarian for many years and then became a lawyer, an employment lawyer, and were appointed to the legislature a few years ago and then ran and won your seat. What brought you to the housing issue specifically?
Megan Egbert: Yeah, it was actually during the pandemic, myself, like many other Idahoans, were faced with no childcare while we were still trying to work. Which was a very frustrating experience that I know many can relate to. And, I have the, you know, fortunate ability of having a mother-in-law, father-in-law who live out of state, who are retired, who were willing to look for a home here to help with childcare during this time. And I think that that was the first eye-opening experience for me was, you know, they were moving from Chicago, or in theory, were willing to move from Chicago- to Idaho to be close to family. And there was just no options. It was not even a possibility to move here during that time. And then, you know, the second thing that has happened is I have older children. My oldest is 20 years old. She works in the medical field. She works at a hospital that’s about, you know, a 20-minute drive from our house, and she recently, she’s still living at home, started looking for a studio apartment, you know, in a, in a suburb of our capital city here, Boise, and cannot find a studio apartment, that costs less than my mortgage for a 4-bedroom house. And so I think there’s just 2 examples of why this, you know, became so important to me. And then mostly it’s just because this is the most important issue facing Idahoans right now. We know this from multiple surveys and just from our own conversations with people, is that Idahoans are getting priced out of living in Idaho.
David Roberts: Mm-hmm. Well, let me stick with you for a second, Megan. I wanna hear a little bit about this committee, this interim housing study committee, you were on. This was modeled on a similar committee they set up in Montana. I’m sort of curious, you had this room, you have these free market North Idaho Republicans, you got Boise Democrats, you got realtors, home builders, a developer, and the apartment association, the city of Boise. It kinda sounds like the setup for a joke. I’m sort of... I’m very curious how that went down in the room, how that felt in the room, and sort of like what you learned through that experience that you didn’t know going in.
Megan Egbert: Yeah. It was an amazing experience. I will say, though, I came to it late. I took somebody’s spot on the committee, when they stepped down from the legislature, so I was trying to do double duty of catching up. I think the beauty of the committee comes in… is exactly though, what you were just explaining of, you know, how different the perspectives were that were on it. And you know, I myself, although, am technically a Boise Democrat, I probably find myself agreeing more with the free market North Idahoans- uh, on many, many issues, this being one of them probably. And I was mentioning before we started here, I grew up in North Idaho. And it was eye-opening for me, I think, that we got to travel the state with this one perspective that we were looking at, right? So we have industry folks, we have legislators, we have cities, counties, we have representatives from everywhere, and we got to travel the state and look at this one problem. And I think the most eye-opening part for me was that it became very clear that this is not a Boise problem. This is not a Coeur d’Alene problem. This is a statewide problem, and I think anybody who participated in that committee would have to say the same thing. We can talk about local control all we want. Local control makes a lot of sense when something is simply a local problem. This is not.
David Roberts: And here’s Seth Cutler of IONNA, the charging network owned by eight automakers, on why one of their sites is technically a bakery.
Final question about the sort of like amenities thing is, are these customized at all regionally or like based on whether it’s in a city or on a highway somewhere? Like, I guess how much work is going into sort of customizing the experience on a location-based way?
Seth Cutler: Yeah, I mean, I think that’s actually the fun part. So, you know early on we said, we said, “Listen, we know we wanna do things differently. We wanna buy land. We wanna build our own amenities.” And in some cases, we went and bought a green field piece of land. It was a, you know, a piece of dirt or grass, right? And we, and we built... And our Houston, our Houston Rechargery is just that, where we built from the ground up an amenity, and we customized it for that location.
What could we fit there? We wanted to build no less than eight charging bays. What could we fit?
David Roberts: Right
Seth Cutler: Well, what could all we could fit was two bathrooms, outdoor seating, vending machines. That’s what we could fit, and we, and it’s a great site, and we know that ‘cause of how much people, how many people come on a daily basis.
But then we went and we found other sites. Hey, this is a, you know, we found in our Apex, North Carolina location, which is just outside Raleigh, the suburb to Raleigh. We cut our first ribbon there. That was a 100-year-old gas station. We thought, “Well, that’d be real, that’d be really cool, right? Let’s take a 100-year-old gas station and make it the service station of the future,” right?
David Roberts: It was closed down?
Seth Cutler: It was, yeah, it was clo— I mean, it was a service station that was opened in the early 1900s, shut down in the ‘60s or ‘70s, and became a I think they were selling kitchen granite top... right, and when we took it over.
So that was, that was unique ‘cause we made, we made that into an experience. That one has a local entrepreneur that sells baked goods out of that site. Right? And the reason for that was because the zoning requirements meant that you couldn’t have a refueling center, but you could have a local restaurant or bakery. So that’s, so it’s classified as just that. And—
David Roberts: Oh, so it’s technically a bakery with chargers attached.
Seth Cutler: Exactly. Yep, 100%, right? So that, so what, we’ve been very flexible as we work with cities and towns, and very flexible with the types of sites that we’re buying or leasing to, again, bring really great charging to customers. And we’ve done that in, you know, there’s like, there’s a list of sites I can give you. Each one has a story of an old gas station of an old natural gas station, of an old car dealership of an old Subway’s restaurant that we’ve taken over, and we’ve converted it to an IONNA Rechargery.
David Roberts: Interesting. The chargers themselves, you’re working with eight automakers. I’m sure they all have their own apps. They all have their own gizmos. How do you make the, like, the literally the plugging in and charging itself, how do you make that consistent across all those apps, across all those different kinds of vehicles, across the two basic kinds of hookups? What does the charger look like, and what does the software look like, and what is the experience like? Am I scanning a card? Am I tapping a card? Am I... You know, is it, is this something where it recognizes the car when I pull up and I don’t have to do anything? What is the experience of charging itself like?
Seth Cutler: Yeah. Well, I’ll start with the one thing you don’t do is download an IONNA app because there is no IONNA app.
David Roberts: Oh, really? There’s no separate app.
Seth Cutler: There’s no app, you know, and there’s, and there’s no intention to make a mobile app. And the reason for that is because, to your point, every OEM, every automaker has an app today. ChargePoint has an app. Presto has an app. EV Connect has... I can go on and on. So we actually have 14 different apps that you can use at IONNA, but none of them are IONNA.
David Roberts: Hmm.
Seth Cutler: And on top of that, every charging station has a credit card reader on it. And we’ve got, you know, Plug&Charge and AutoCharge with, you know, a significant number of OEMs, even beyond just our eight, right? So we have the eight investors that we’ve turned on Plug&Charge for most of them, the rest coming very soon. But we’ve also turned on—
David Roberts: So that means— So just to, just to clarify for listeners, that means I can drive up, grab the charger, stick it in my car, and it recognizes my car and will charge it to my account without my scanning anything or tapping any, or doing anything.
David Roberts: From there to Uruguay. Ramon Mendez Galain, who I’ve been wanting to talk to for years, ran that country’s transition to 98% renewables. Here he is on how he got every political party to sign on to the same plan.
Well, one of the most sort of fascinating, at least from where I’m standing here in the United States, sort of baffling aspects of all this is you came in a majority government. You had a political majority, but the president pushed you to basically generate political unity on this, to create a plan that every political party would sign off on, basically, so that these changes would be enduring and steady over time. And of course, I’m sure any country would love to have political unity to make their plans enduring over time. But how in the world did you get everyone from every side of politics to sign off on these very ambitious plans?
Ramon Galain: Well, this was one of the two or three most important issues that allows us to make this transition, because, yes, what is happening for the last 20 years in our Western democracies is that we’re going from one government to another government, and things change dramatically.
David Roberts: Yes.
Ramon Galain: And even, it’s even worse because the energy sector is, in a way, politicized and ideologized, which means that you put the ideology in that which is a technical and economical environmental discussion-
David Roberts: Mm...
Ramon Galain: which in fact has no meaning at all. So what I did is when the president asked me to lead a technical and political committee trying to figure out a long-term solution or long-term transition or a new framework for the energy sector in the country, my proposal was, let’s try to discuss what the benefits are for our constituencies, no matter if they vote right or left or center or whatever. I mean, what are the benefits of going to one direction or the other? And the crucial point was, let us not begin discussing about what we should do today, not even next year, not even during this government. Let’s try to imagine how we can think the country should be in 30 years from now.
David Roberts: Mm.
Ramon Galain: So we are thinking about 30 years in the future, everything becomes much more simple when we’re discussing. Because if you’re asking about what we should do now or next year, the o- your first thoughts are, and what am I gaining or losing in this direction? Should I have more and less power? Should I have more and less votes? Should I have more and less money?
But if you’re discussing about the future, the real future, you don’t imagine yourself 30 years in the future. I mean, it’s simple to discuss how you can imagine the country if you’re not in this imagination. So it was not that difficult to imagine the country that we would like to have 30 years in the future concerning the power sector, and then we make a kind of reverse engineering. So if we have to be in 30, you have to be in this point, where should we be in 25 years?
And when in 20, and when in 15? And then when we’re going back and back, it was becoming a little bit more difficult. But at the end of the day, if we have bo- want to face on this imagined future, then we have to take decision now. And this was the way that we built an overall agreement. But once again, the most important thing was that we build it according to the benefits of the new system we bring to our citizens, and this was the crucial issue.
David Roberts: Well, speaking of that, I’m curious what role, if any, climate change played in that discussion. Was that a big part of the puzzle, or were you focused more on just basic financial and security benefits. What was the sort of discussion like?
Ramon Galain: It was one of the most important point, but not the most important one.
David Roberts: Mm.
Ramon Galain: The point at that time, and now it’s much, much simpler, but at that time the point is that for any country, but particularly for a country, oil imported country, a gas import country, the most important was try to have the basis of our system based on our own national resources. Because if you have your own national resources, you don’t have to depend on fluctuation of energy commodity prices-
David Roberts: Right.
Ramon Galain: on decision of wars or whatever decision that you cannot take, but other takes for you, and you have to pay for the consequences of that. And renewables for us was for sure the best option, because renewables, you are not dependent on an energy commodity price. You are much less vulnerable to these geopolitical events all over the world and are your own national resources. So you are much more sovereign and much more independent, much more stabilized as an energy system.
David Roberts: Also this month, Quinn Nakayama of PG&E on the question I end up asking every utility eventually, which is, why would you ever adopt something that saves money?
Well, this brings me to my final question. Quinn, maybe this is where you’re going, but this is probably the question that my listeners have been waiting for me to ask this whole hour, which is, if the utility spends a capital dollar on a pole and a wire, they spend a large chunk of money, and they get a guaranteed rate of return on it, they make money that way. Whereas if you come in with some digital solution and you say, “Here’s a super cheap software-based solution that can help you avoid the need for that pole and that wire. Look, you can save money,” a normal business would be like, “Oh, good. I get to save money.”
But a utility business, if they save that money, makes less money. You know, this is something we come back to again and again and again on this pod. The utility wants to spend money. That’s how they make money. And almost everything we’re talking about under the heading of innovation is, one way or another, something that avoids the need to spend a bunch of CapEx on big infrastructure, i.e. something that is going to reduce an investor-owned utility’s profits.
And I just don’t know, it just seems to me like the very basic business model we’re talking about here, the very basic regulatory structure of these utilities is working against innovation almost intrinsically. How do you get around that? I guess is my question. Like, how do you work around that very basic mismatch of incentives?
Quinn Nakayama: I would say there’s a couple of things here. One is that used to be the case, and I would say that, well, if you take a look at a lot of the utilities that had 0 to 2% flat growth in their utility sector, and they weren’t building a lot of infrastructure, and all of them would be like, “Yeah, let’s do more capital. We need to figure out how to get our rate of return.” That may have been the case in a lot of utilities in the past.
Now, all these utilities, including ourselves, are like, “We need to get rid of some of this capital ‘cause we can’t spend it all without really ballooning rates.” Like, if there are ways that we can defer capital spend, if there are ways that we can have alternative ways.
David Roberts: It may be in California, Quinn, but look out at the country. Like, I see utility executives just, like, drunk on this. They’re like, “Heck yeah. Like, you want us to build a lot? We’re gonna build all we can build. We’re super excited to spend CapEx. We’re super to spend more and more.” Like, And I’m not sure that the political blowback has fully reached them yet, or, like, changed their mindset yet.
Quinn Nakayama: I would say there was a swing where all of these data centers happened very quickly that caused all this infrastructure to build, that be built out, that then caused this lag on the generation side. And, you know, more broadly, it was just, it just happened a lot faster than a lot of folks could react to. But if you think that this whole politics on raising rates and the blowout that’s happening on rates side isn’t on the top of executives’ minds across the entire nation, I think that’s just not true.
And I think the reason why is that, listen, you know, utilities actually don’t want legislation coming into their area. I don’t think that they actually, utilities writ large, because you have everybody commenting into areas that maybe people are not as well informed into. And the utility space is very complicated. It’s an engineering nightmare. It’s, you gotta think about protection, you gotta think about a whole bunch of other type of things.
And so I would say that most utilities out there, most utility executives are very concerned about customer affordability. They just need to figure out how exactly do they incorporate all of this AI data center-related large growth that’s happening super fast without coming across as being obstructionist toward that growth, and at the same time figuring out how to do this with less capital. I feel like, you know, that pendulum is vastly swinging the other direction now.
David Roberts: Hmm.
Quinn Nakayama: We have way too much capital everywhere. And wildfire, by the way, is not just a California problem. I don’t know if you’ve seen, but there’s wildfires happening everywhere.
David Roberts: Yes, I noticed.
Quinn Nakayama: So, like, hardening infrastructure for capital, everybody’s gonna be coming out of their ears with capital. Now it’s gonna be about how do you do, how do you meet these objectives that you have on these really big ambitions that the state has on load growth, the state might have on wildfire-related risk or whatever other catastrophes that they’re having with the changed climate that’s already out there, and then trying to do all of that with less money because the rate blowback is just gonna be too big. I think that in order for us to do this effectively, though I think we need to figure out how to take some of these O&M-related expenses-
David Roberts: Operation and management.
Quinn Nakayama: operation and management-related expenses and figure out how to peanut butter them a little bit more across a broader range of time, which smells like capital, but what I’m just trying to say is that when you have an operations and maintenance-related expense, it hits the ratepayer the next year, right?
David Roberts: Yeah.
Quinn Nakayama: A capital project that’s $300 million that’s spread over 40 years hits the ratepayer much less over a long period of time. That’s why I think even in a utility such as California, we don’t want a lot of O&M expenses on our books just because next year that’s gonna hit your rates. And a capital project can spend a lot more money, but hit the ratepayers very fractionally as ratemaking occurs in all of our areas. So that’s just something to consider. It doesn’t really necessarily have to do with a rate of return.
David Roberts: After that, Varun Sivarim of Emerald AI on making data centers flex down for the grid and who eats it when that goes wrong.
I have some questions about just the computing front. So one of the things your framework has is these tiers of computing workloads. There are jobs that can’t be touched, that have to run all the time. There are jobs that can tolerate 10% throughput loss, 25%, 50%, there’s these tiers. So one of my questions is, who decides what’s in what tier? Is that something that the owners of the compute tell you, or is that something that Conductor analyzes the compute and makes its own judgments about what is and isn’t movable and controllable? Where do the tiers come from?
Varun Sivaram: That’s a great question, Dave. It’s the heart of what we’re doing, which is serving both sides at the same time, the energy grid’s needs and the AI customer’s needs.
Look, at the end of the day, AI customers are doing the most economically important work in America, right? Generating valuable tokens of artificial intelligence, and you don’t wanna mess with that in a way that’s going to degrade their expected quality of service. And so yes, today, for our commercial scale deployments, we receive tags from the customers, the users of AI compute, tagging which of their jobs has which flexibility parameter.
And it’s more than just, you know, this can be flexed and this cannot be. Some jobs, a benchmarking run, for example, you really can’t downclock the clock frequency of the GPU ’cause it’ll mess with the benchmarking run. Some jobs, a fine-tuning run, they can be deferred. They could also be downclocked. But some particular workloads, like a serving inference job, really can’t be downclocked or paused because the user expects an answer within a latency bound, and perhaps that needs to be geo-shifted.
So within different priority tiers, you know, low, medium, high, or on a scale of 0 to 100, as well as across different characteristics, preemptibility, throttle ability, the ability to be paused and checkpointed, or the ability to be moved within residency requirements. It’s a complicated set of constraints, and we feed this in through what we call the normalizer to turn it into actions that we can take, and then our AI agents take the actions.
David Roberts: Well, I’m sort of-- Well, I mean, one of the things I wonder is if somebody gets that wrong, who’s responsible? Like, who’s holding the bag at the end of the day? Like, if you could downclock the power, a job slows down, and for some reason that someone didn’t anticipate, that breaks a job or doesn’t get the result it wants, or there’s an unanticipated result, is that just on the compute load for mislabeling its job, or who’s responsible if something goes wrong with the jobs?
Varun Sivaram: Yeah, great question, Dave, and I’ll use this to mention, you know, this is high stakes, high reward in kind of three different dimensions. Number one, it’s massively high reward because if we make this all work, we can fit 100 gigawatts, as you mentioned from the Tyler Norris paper, we can fit 100 gigawatts of AI data centers onto the grid that we didn’t think we could fit, and we’ll walk through the benefits of that.
So massive reward. High stakes though is on both sides. If you don’t do this right, the grid could face reliability issues. On the other side, if you don’t do this right, the AI customers might face failed expectations, again, for the most economically valuable work in America. And the third dimension, of course, is a security dimension.
AI data centers becoming energy technologies and this fusion of AI and the energy system where we sit at the interface is a potential security threat, right? These are two massive multi-trillion dollar networks that are critical infrastructure for America’s economy. It’s the backbone of our future economy. And so we gotta make sure that we’re doing all of this in a very safe way that protects the reliability of the grid, is resilient to intruders or threats from abroad and domestically, and protects the integrity of the AI workloads.
Your specific question, Dave, is who pays if there’s a screw-up? And the answer to that specific question lies in the contracting, right? There will be, depending on the context, the grid, if they’re offering you a faster connection, they may say, “You know what? If you don’t curtail your power when we ask you to, there may be a financial penalty, or there may even be a technical penalty. I’ll flip your breaker.”
David Roberts: And finally, Trump used the Defense Production Act, a 1941 war powers law, to force a ruptured California oil pipeline back online for one of his donors. Todd Tucker of the Roosevelt Institute discusses whether the next administration should hand that tool back.
Run us through the Sable Case, just so we can have the context for what we’re talking about here. Just briefly talk about the role of the DPA in that case.
Todd Tucker: Yeah. So as you established in your introduction, this is a case that has been kicking around, you know, California courts for years now, and a company, Sable. Sable is a Texas company that was involved in the offshore oil industry in California, and they had been tied up in litigation in California, you know, for about the last decade, trying to get approval to restart production out there. When Trump comes back into office, they start floating the idea of using this sort of FDR-era tool, you know, Defense Production Act, which had its roots in the second War Powers Act of 1941 under the Roosevelt administration, that gives government all this power to not only prioritize governmental contracts, but actually to reorient the structure of contracts, even in private markets.
And so that’s what Trump did here. He basically said, “Any companies that can help Sable produce oil in California need to prioritize contracts for Sable in order to get that going as quickly as possible.” And you know, this is something that the DPA is a tool that the Pentagon uses a lot for defense and national security purposes, but this is one of the more ambitious uses, again, sort of really going back to, you know, the mid-20th century in doing it for a non-national security purpose, really just more to benefit an individual fossil producer. It will have some impact on oil prices, which is ostensibly sort of a nexus here that merits DPA use, but not very much. I mean, it really is more of a handout to one of the president’s campaign contributors.
And so, you know what… The argument that we make in several Roosevelt pieces is that this precedent has now been set to use the DPA for clean as well. If it can be used for fossil for these purposes, then it can be used in a even smarter fashion for clean purposes. And you know, I think that, for instance, some of the hiccups in the offshore wind industry, sort of around 2023, would’ve looked very different if the federal government was stepping in to actually solve some of the supply chain and other challenges that Orsted and some of these companies had.
David Roberts: I mean, the question I wanted to ask is that the provenance of the DPA is sort of contested. It’s in court right now. There was a, you know, one judge said, basically, “Yes, you can use it to override California law and regulation,” but that’s gonna go to the Ninth Circuit, and there’s some speculation in your volume and elsewhere that if that gets affirmed, if a judge says yes, the DPA can override state law, that the Trump administration is just kinda gonna go on a bender overriding all kinds of California state laws and other state laws in other states, doing whatever it wants on the basis of the DPA.
And so I guess I would just sort of ask flatly, do you want the Ninth Circuit to uphold this judgment and affirm that the DPA can override state policy? Like, if you had your druthers, is this an affirmative good, or are you just talking about making the best out of something bad that’s happened? Do you know what I mean? Like, would you actually want the DPA to be used this way and to be affirmed as capable of being used this way?
Todd Tucker: It’s a totally fair question, and I am of two minds of it. You know, I think on the one hand, in the very near term, it is obviously troubling the way that the Trump administration is using not only the DPA, but earlier the International Economic Emergency Powers Act to impose tariffs. You know, he’s using a lot of the best of the New Deal legacy to hammer home things that are sort of narrow and shortsighted and to benefit his campaign donors. You know, as someone who’s worked a long time at the Roosevelt Institute, it gives me no joy at all that it’s Trump rather than Biden that is using these authorities to the maximum.
On the other hand, in part because of all of these problems and weaknesses in our democracy and the fact that, you know, a single senator from a single state can block all sorts of a policy from going forward and that, you know, filibuster rules make it really hard to do anything other than tax credits, right? We have all these kind of fundamental problems in our system, and tools like the DPA or the International Emergency Economic Powers Act give the executive branch a way to move forward even when Congress is being obstinate.
So, you know, I think what we argue for in this volume is that if these tools are gonna be on the table, we need to get rid of this asymmetry where, you know, right-wing anti-climate governments use their powers to the hilt, but when progressive governments are in power, they sort of hold themselves back in the name of these outdated norms.
