Turning up the heat on utility costs

T.J. Thornton (00:00): Hello and welcome to Global Research Unlocked, where we discuss what's rising from growth industries to rising risks and opportunities in global markets. I'm T.J. Thornton, Head of Product Marketing at BofA Global Research, and we're recording this episode on Monday, January 12, 2026.

Ross Fowler (00:20): What's happened in the second wave of data center tariffs, and this has happened in areas like Georgia. Indiana and Iowa are doing it now, where they actually designed the tariff to charge above the data center's cost of service. They're making the data centers pay more.

T.J. Thornton (00:40): Utility stocks, especially generating assets, had a good 2025, but the landscape in 2026 is a bit different. Affordability proved that it can impact elections, so it'll be a big focus this year, especially with midterms and gubernatorial elections. That could impact utilities and data center development in theory. Many of the generating companies did have a good year in 2025, so what's left and what should we read into the recent underperformance of some of the stocks? We're joined today by Ross Fowler, Head of North America Power and Regulated Utilities Equity Research. Thank you, Ross for joining.

Ross Fowler (01:15): Hey T.J. Thanks for having me on.

T.J. Thornton (01:18): Sure. affordability, I mentioned that's one of your key themes for 2026. 36 states have gubernatorial elections this year. Of course, it's also a midterm year. What does that actually mean for utilities and do you think it'll be a challenge for regulated utility stocks?

Ross Fowler (01:36): Yes it's definitely a big theme and we wrote about this in our year ahead, but particularly for the regulated utility stocks. Remember that all utilities are economically regulated at the state level. There's not a lot here that the Federal government can do around affordability, but it certainly will be an issue du jour for the midterm elections. As you think about it, it might be that electricity bills are the new exit. Certainly, led to a lot of consternation in the New Jersey elections and in the Virginia elections which were the two off-cycle elections we had in 2025. I think as you look at 2026 and the way the year lays out you've got to think about the states that are really going to have an affordability issue and a contentious election. There's several states where you have utility rate cases that are active in those states, but you really should probably focus on the primaries because they've leaned so far politically, one way or the other. It's generally probably not a November issue around the general election. It's more an issue for the primary. Where you probably are more concerned around election issues and affordability becoming a bigger issue are states where you have active rate cases for utilities, and it's much more likely to be contentious in the general election because it's a swing state. Those are Arizona, Michigan and Pennsylvania. The primaries there respectively are August for Arizona, August for Michigan, and May for Pennsylvania but that could be a full year issue right through to the November elections. Interestingly some of those commissions that determine economic regulation at the state level are actually elected. Not very many, about two thirds to three quarters, are actually appointed by the governor in the state, but there are a few that are elected. There are two seats up at commissions in Alabama, Arizona, Georgia, Louisiana, Montana and Nebraska. We would watch those elections very closely. That's absolutely an issue. It plays to both sides of the political aisle. It's not going to be as simple as 2025, where it was all about growth, demand, AI, data centers, everything's up and to the right. It's going to be a little bit different, as you said in your intro. You're going to have to be more nimble. It's going to be around stock picking and some of this political headwind is going to move in and out depending on the primaries in the general election.

T.J. Thornton (03:56): Okay and looking broadly, you mentioned AI, of course everybody's favorite theme, sorry to say, but there are some investors that care more about AI than regulated utilities. We'll talk about what all this could mean for that story. Do you think affordability issues, because they are often associated with data centers, come to town and suddenly my bills go up. Do you think that could slow down data center builds by limiting the amount of power that can be developed that these data centers can get?

Ross Fowler (04:25): The answer to this question is really it depends on where you are. To real estate, it's all about location. Data centers are generally clustering in areas where they have access to the resources they need. That's not only power, it's also land, it's also cool the data center. It's also now becoming political palatability. What we're hearing when we talk to our CEOs is one of the questions, the hyperscalers and data centers are asking, is if I go into Community X, am I going to get a lot of pushback in putting my data center there? The new 800 DC volt architecture really increases power density. You're going to go to about a megawatt per rack, that's a lot of power in one rack that you're going to put into a huge building and they want to go where they're welcomed to build these giant data centers that are becoming more power intensive. The current clusters are generally speaking in California, around Silicon Valley. I guess that's not surprising. Northern Virginia where the NSA and the CIA are where believe it or not about 70% of the world traffic travels through, not just the United States is, Texas and the corridor between Milwaukee and Chicago in Wisconsin and Illinois. What we see is you're going to stay in those clusters, or to the extent you start running out of power availability in those clusters, you move to the next nearest clusters. Where that's happening and where the permitting is happening for the next data centers to come. Arizona is the next cluster moving from that California side. Texas is just more of Texas, in that Virginia, Northern Virginia cluster, they're spreading into Pennsylvania and then Iowa and Indiana in that Chicago Milwaukee corridor. It's really important and you can look at our year ahead for this as well, what the tariff structure is in your state. The first derivative of data center tariffs were just what utilities have always historically done. Just charge you your cost of service. Now they did modify that a little bit as data centers came in and said, okay you've got to pay that rate for your electricity, but you also have to give us some upfront funding for the capital needs of building power plants or building transmission lines that you need. These are constructs that have appeared in Louisiana, Mississippi, some other places across the south, Virginia, that Northern Virginia pocket. What you worry about there is under the hood, there might still be some cross subsidization, as the grid becomes constrained and you need to move away from the data center to maintain reliability and or even if the data center increases volumes massively for electricity and spread your fixed costs out over more volumes, and that actually lowers bills. What's happened in the second wave of data center tariffs, and this has happened in areas like Georgia. Indiana and Iowa are doing it now, where they actually designed the tariff to charge above the data center's cost of service. They're making the data centers pay more. That becomes interesting because now if you have an affordability argument that you're trying to fight, you have an auditable, provable subsidization of residential customer bills and you can outright argue very clearly and definitively that, hey, the more data centers that actually come into this service territory, the more your bills go down, the more subsidization you get across. One last point on this, the deregulated markets in Texas, markets where generation is deregulated in Texas and in what's called PJM, which is Illinois over through the Mid-Atlantic. Texas is already unlocked because you can build stuff very easily in Texas. Data centers are moving in there aggressively. PJM was not unlocked in any real way because there was a lot of consternation as to how the grid could or would be constrained should you add a lot of data centers to it. The FERC or the Federal Energy Regulatory Commission in mid-December ruled on this colocation issue. Colocation is just putting a data center next to an existing power plant. When those tariffs are finally finalized in April, and even probably beforehand, you're going to unlock a lot of data center deals. In fact, we got the first one just last week. We think PJM is ripe for data centers moving in, and then the other cluster window where we're going to see things happen in 2026.

T.J. Thornton (08:50): Okay, got it. A quick follow up on that. Do you think that data centers are generally okay with paying these high rates? I guess it depends on where they are. We hear about lots of them going into Indiana. I think you said that they have that sort of tariff structure. Have they still been willing to sign up given those terms?

Ross Fowler (09:08): Remember every state has a different price of electricity based on the generation stack that's embedded in it. If you think about Illinois and up to that Milwaukee corridor, that's very cheap power because that's a lot of the net legacy nuclear fleet of the country. Indiana has some residual old coal resources, probably unsurprisingly, and a lot of wind that's put in the meantime where power's pretty cheap. And Iowa has ridiculously cheap power because it's about 60% wind resource. A lot of wind blows a lot of the time. Versus other places, the base rate you're starting with is relatively inexpensive versus a lot of other locations. Then if you charge a little bit more to the data center, you're still highly competitive with alternatives. Ultimately, data centers are not really worried about what they're paying for power. It's a huge cost of running the data center but it's not a huge cost for them as they build out the data center. They're desperate for it. Right now, all of the negotiating leverage in some of these conversations or all of these conversations is on the power utility side of the table. They need it and they're willing to pay up for it.

T.J. Thornton (10:17): That makes sense. Another one of the things that you talked about in the Year Ahead was buying weakness in some of the power stocks. I'm frankly a little bit surprised that they've been weak. Of course, some of the AI stocks have come off. We'll need power for the next decade and it's going to grow at a clip much faster than where it's been growing. Why have these names been weak?

Ross Fowler (10:37): The weakness really started around about Thanksgiving. People came back from the holiday and just decided to sell winners. Some of those are in those AI baskets. When that happens, correlations go to one, in some regards. We think that's the market not really differentiating between the AI players, where there have been some emerging concerns about leverage that have been out there. AI is where that spending is going. On the tech side, you could theoretically worry about spend that's been out there in some of the press and the articles that I've read. You want to be on the receiving end of that spending because I don't think they're slowing down. That's certainly not what we're hearing from our conversations with both regulated utilities and power companies is that spending still coming, interest is as high as they have ever seen it. That was said by utilities CEO’s to a man at the EEI conference in November. I think you'd got wound up in the calendar and protection of alpha and the wind down. The FERC co-location order in mid-December was not as impactful as it should have been to the market because of when it happened. As the market catches up with this , as they take a fresh sheet of paper into 2026, as you get the next deal beyond the one that we saw next week. People will get back into this trade and understand how the environment is from here.

T.J. Thornton (12:02): Okay. Dimple's been on the podcast. We've talked about sources of generation. I want to talk to you about that as well. We know there will be growth in power demand or we're pretty sure of that. Certain sources of generation are being taken off like coal. What do you think the mix of new generation will look like when you think about solar versus nuclear and some of the other options?

Ross Fowler (12:23): It's really a progression of what's needed. As you think about electricity demand, remember it's very weather dependent and it's very time dependent. Think about it in two ways. On a regular average day, power demand is very high in the morning because we're all getting out of bed and turning everything on and getting ready to go to work. Then it fades when we're all at work and there's not much power usage on the residential side at home. Then it peaks into the evening when we all go home, turn everything on again. There's also seasonal variance in demand, just depending on the weather. There was no demand growth on the grid between 2010 and 2020. After COVID and the pandemic, the first thing to hit the grid was reindustrialization, reshoring of supply chain , and now you've added AI demand on top of that. The grid isn't really ready for the demands that's come to it. You're dealing with highly capital intensive projects, long dated to build things. The grid's going to take a while to catch up to this from the supply side, if this demand continues and we believe it will. On the near term, what you're really focused on is that peak in the summer and that peak in the winter. Because you can't store electricity, you have excess capacity and all those other times of day and all those other times of year right now anyway. As you manage that peak, the two things that need to come onto the grid are solar and storage, which helps you massively in that summer peak. You can use that solar resource and then if you have storage attached to it, it can extend into that peak after the sun goes down when we're all at home, before we go to bed. Then you need simple cycle gas turbines. You can turn them on very quickly, like starting your car, ICE engines, reciprocating engines, things like that can help you very much in that winter peak in the morning when it's cold because the sun's probably not out then. Then as you play this demand growth out over the next few years, you start to run out of that excess at other times of year and other times of day. Now solar and storage doesn't really do much for you. Reciprocating engines don't do much for you. You need big dispatchable base load power, and that's only two sources. It's going to be big gas plants, combined cycle gas turbines. But I think you're going to see that come and then you're going to see eventually nuclear, although nuclear takes a long time to build. If you think about the progression, it's solar storage, reciprocating ICE engines for now, for the next year, couple years, then you move into big gas plants. Then maybe in the 2030s you add nuclear to that equation.

T.J. Thornton (14:48): Quantum computing. Jumping around here on topics. Actually just earlier I was talking to our hardware analyst who's been focused on this. He's got a primer out on quantum. One of the promises of quantum is that it can process much more efficiently than classical computing. You just need a lot less power in order to do the same amount of processing. Is this something that people are at all focused on? The fact that quantum may be coming down the pike and what that could do for power consumption?

Ross Fowler (15:18): It's very theoretical, as you said, kind of like fusion on the demand side, although fusion may frankly take longer than quantum computing. It is something that comes up in conversation but it comes up in a broader context about will this AI demand all slow down and go away? Then the secular bull case is gone. I don't see that right now, but people are worried about the tail risk of another deep seek moment or the recent cooling advancements that we saw more recently. The cooling advancements themselves though, all you're going to do is take your power use effectiveness which is how much power you're actually feeding to the chips and actually increase that or get that closer to one for one. Right now it's at two and a half , so a lot of the power you are using is going to cooling. You'd rather use all that power for the chip processing. Quantum computing uses way less electricity and even in combination it could lower data center demand by about 12.5%. Full use would require the power of like heating water for your afternoon tea to run a data center. It's crazy effective from that regard. However, there's massive problems with it. It's very noisy. It's very error prone. It's really hard to imagine that there's a lot of AI applications that's acceptable for. If you could control that, get lower error out, then it is really a thing but that's going to take a long time from my perspective. It's going to move into that 2035 or later period where we've already built out a bunch of data centers and then, maybe it all has to be rebuilt to quantum computers. That quantum mechanic issue is very hard to control. It's one of these left field things like fusion that if it happens next week and somebody has a massive breakthrough, then either quantum computing from the demand side or something like fusion from the supply side could create a headwind to the secular bull growth story. The probability of it right now at least seems very low.

T.J. Thornton (17:12): You mentioned some of the supply deals. We got another one I think just last week. 20 year deal between one of your power producers and a mega cap tech stock. This is not the development of new generation, it's for an increasing output from existing generation. I think this one's mostly nuclear. But how many more of these deals are there? How much spare capacity is there in power that could then be sold to data center customers?

Ross Fowler (17:39): One of the concerns you've seen with this and I think there's a lot more. You go back to the FERC co-location decision we were talking about from mid-December. Everybody's been worried about affordability in PJM. Capacity prices have gone up massively because of data center demand. People's power bills are up 20 to 30%. If you're in Pennsylvania, you actually have excess generating capacity for electricity. You export it all to New Jersey, Maryland, and states to your east. What Pennsylvania seems to be adopting as an energy policy is no I want to use the electrons I generate in my state for me, for data for my own tax base, for my own economic development. Now that's not very good for Maryland, Delaware, New Jersey, and the states out to the east of Pennsylvania. I don't think Pennsylvania sees that as Pennsylvania's problem. Illinois is very much the same way. That's frankly where most of the nuclear fleet is. There's a lot more nuclear deals to come. No real need for what's called additionality, where you bring more megawatts to the equation as you take these nuclear plants and commit them to data centers. Now, I think politically you probably still want to be very careful as you design these deals. The one we saw last week involves upgrades of the nuclear plants. There's some of what we call soft additionality coming in connection to that. You don't need full additionality to do these deals. Look at some point, do you run into a reliability problem on the grid? Certainly. But that's probably a '27 or '28 issue as you continue to sign these up. Nuclear will be the first. Nuclear is the easiest. One because you have the things you need there. You have a lot of land next to a nuclear plant. You have a lot of water to cool it. And the amperage and voltage that comes off of a nuclear plant is far more stable, so you need less equipment on the data center side to stabilize that over time. Once a nuclear plant's on, it's on, it's running 24/ 7 until it's two year refueling outage. You don't have to worry about maintenance cycles and so forth like you do on the gas side. Eventually we'll get to gas deals because you have to. There aren't that many nuclear plants in the country, if all this demand continues to come. The issue with gas deals is you do have two maintenance cycles a year where the plant's down for a month or more. You'll need more gas plants to feed the same data center capacity. Somebody's got to take the natural gas commodity risk, whereas uranium prices have been fairly stable.

T.J. Thornton (20:01): I suppose all that excess generation capacity that they have in Pennsylvania maybe that's solace to the Eagles fans out there after the loss yesterday.

Ross Fowler (20:08): It's certainly not good if you live in New Jersey.

T.J. Thornton (20:10): Especially if you're in New Jersey and an Eagles fan, then it's double jeopardy. That is the upshot of this. Maybe it's decent for Pennsylvania, but then other jurisdictions could still wind up paying higher power prices based on deals like the ones we were just talking about.

Ross Fowler (20:25): At some point there's probably a political resistance feedback loop to that, but it's very hard in a competitive market. I see a path to political and affordability resistance in a regulated market because remember, generation through the wire to your house is all regulated by the state at that point. If you're not in one of these places where it's going to be accepted or you've set up a tariff the right way, you can get a lot of pushback and you can win the argument in some sense because you have to just get that state regulator or governor on your side. If you're in a deregulated market like Texas or PJM, the generation market is deregulated to most degrees. If Pennsylvania says we're going to put more data centers in because that's good for us, and it raises the power prices or the utility bills that are passing through to other states, it's a massive conflict of interest. Pennsylvania wants it one way and says I want economic development, so does Illinois and the rest of PJM is going to suffer for those decisions. There's no way for PJM to come back to Pennsylvania and Illinois outside of saying reliability of the grid is constrained and stop them from doing they're doing. At some point you get there, but not today. Certainly not through the next several nuclear deals.

T.J. Thornton (21:43): Switching gears one last time, last question is about wildfires. California utilities have been under pressure because of litigation around wildfires. The fires around LA were almost exactly a year ago. There were fires in Colorado a few years ago that was also blamed on the utility. How do you think this all plays out for utilities and consumers in these states with big fire risk? Does it just necessitate more spending to reduce these risks and that means higher bills and a tougher regulatory environment? Generally do you prefer to just stay away from utilities in such states?

Ross Fowler (22:15): It really depends on a couple things. It depends on valuation. If the valuation is trading at too steep a discount for what we believe the wildfire risk to be then we are interested in those stocks because we think there is a way forward through all of this. I will say it's interesting a lot of people focus on wildfire risk out west. We think it's a national issue now. We had wildfires in parks in Manhattan up Inwood a couple years ago in the summer. It can happen anywhere. The problem is when is it utility caused and when is it not? Unfortunately, tragically it seems like it was utility caused for the fire in California. We don't know that for sure yet. There's a couple things western states have done outside of California and we're hopeful that California gets through this legislative session in April, and starts to find broader solutions going forward. But states have passed legislation that essentially develop wildfire mitigation plans with the utility through the state regulator. Then as long as you are adhering to that plan , you are predetermined to not be negligent. There's a lot of legislation in places like Utah and Idaho that have come up with structures to say, if you're doing everything that we want you to do for mitigating wildfires, even if your equipment, tragically starts something you're not going to be held to fault now. You've got to work that through the court process. You still have to test it, but that's a very big backstop. Utilities are spending a lot of money to mitigate things. You can do covered conductors, you can put some lines underground. Power system shutoffs when the wind is very strong. All of that is working to better solutions. We think the fire risks are probably overestimated in terms of financial impacts. The only place that we still worry about it is California. Basically because the topography of California is different. If you think about some of these wildfire states like Colorado or Montana or Idaho, there's not a lot of trees. It's very high desert. It's very plains grass. Even the fire we had in Texas was all sort of grassland, grazing land. The value of what can burn on a relative basis as tragic as one of these wildfires is not as high as what can happen in California. In California if you get a wildfire, you can get into, as we saw tragically about a year ago, you can get into, some very suburban city neighborhoods very quickly, and a lot of homes can be destroyed. We would still put a discount on California even if they fix this in April. Overall we think that the wildfire risks are probably overestimated by the market, not underestimated by the market.

T.J. Thornton (24:55): Value opportunity. Thanks Ross for joining us on a range of utility topics today.

Ross Fowler (25:00): Absolutely, T.J., glad to be here.

T.J. Thornton (25:03): As for utility bill affordability, it's something that the states have more control of than the Federal government. Ross thinks the issue will be front loaded this year as it becomes more of a talking point in the primaries than in the general election. One of the ways that states, especially those that are good locations for data centers, can address affordability is by charging higher tariffs to data centers. This is already underway in some areas where consumers and utilities can both win. When it comes to the source of generation, that's complicated by seasonality and timing of overall demand, but base load will be part of this, and yet there are constraints when it comes to new base load generation. There is room for expansion of existing generation though, and wildfire risk seems like a particularly difficult problem for utilities, but there are plans in place that suggest some utilities in fire prone states are actually decent opportunities. Thanks for joining.

(26:07): “Bank of America” and “BofA Securities” are the marketing names for the global banking businesses and global markets businesses [which includes BofA Global Research] of Bank of America Corporation. Lending, derivatives, and other commercial banking activities are performed globally by banking affiliates of Bank of America Corporation, including Bank of America, N.A. Member FDIC. Securities, trading, research, strategic advisory, and other investment banking and markets activities are performed globally by affiliates of Bank of America Corporation, including, in the United States, BofA Securities, Inc. a registered broker-dealer and Member of FINRA and SIPC, and, in other jurisdictions by locally registered entities. ©2026 Bank of America Corporation. All rights reserved.