← Currents
Currents
Ep349: Meeting Surging Power Demand with Gas Assets
00:28:54
John Bills and Adil Sener of PEI Global Partners discuss the renewed interest in gas‑fired power assets driven by rising electricity demand, data center growth and supply constraints in power generation.
Loading summary
Transcript45 lines
- [00:00]
A
Foreign. To Norton Rose Fulbright podcast Today we're recording with John Bills and Adil Sener from PEI Global Partners, which is an investment bank active in the power generation markets. We asked him to join today to talk about the renewed and very heightened interest in gas fired assets. John and Adeel, thanks for recording with us today.
- [00:25]
B
Thanks Todd. Good to be here.
- [00:27]
C
Thanks Dant.
- [00:28]
A
All right, so as I mentioned in the lead in here, five to ten years ago we were doing all kinds of gas fire deal with all these exotic hedges and things and then kind of the market went quiet for several years and now we're in the midst of a lot of M and A activity in the national gas fired generation sector. And we can talk about the financings too. But let's start first with the M and A side. Why all of this renewed interest?
- [00:55]
C
Well look, that's a great question and I think the short answer is that the demand growth really since the Great Recession, when you go back to 2010s, we had almost 50 years of flat demand growth across the country. And during this time we retired bunch of thermal generation and added renewables, mainly solar and wind. And going back to 2010, actually looking at from system planning perspective, when planners actually saw almost zero demand growth in many regions of the country and the directive, when the directive was really to decarbonize, that was the right thing to do, right? You basically, you have an oversupply, you basically add renewables, mainly solar and wind at declining costs and try to replace your generation towards lower intensity emissions. But fast forwarding to today, I think what no one, almost no one saw is that this tremendous amount of data center demand growth coming into the picture over the next five years we are going to add a large European country size demand electricity demand that is going to take roughly 12% up to 12% of the US electricity consumption. We are not prepared for that as a sector because we have not really forecasted this kind of demand growth. And as you know, well, power development takes time. You need turbines, you need human resources, you need siting, you need permits and everything else. The cycle itself is five to seven years to build and develop these baseload generators. And we are in this period where we are seeing the capital is flowing, human resources is flowing and thermal development is restarting and ramping up. But we are in this period right now for the next, I would say for five years. There is scarcity of generation iron on the ground that can provide 24,7 baseloads. That drives up the value of everything iron on the ground. That can generate power. And that's why we are seeing this rush to existing generation.
- [03:14]
B
And I would just add, you know, big picture at the end of the day, the commercial opportunity to own a gas plant is tremendous right now. Whether that's in capacity, revenues, energy revenues, depending upon the market. And now with these long term less price sensitive buyers such as data centers and longer term load that need it, we're seeing contracts being entered into 10, 15 plus years at really attractive prices. And really we haven't seen that Certainly in the 30 years I've been in the business or so it's unprecedented the amount of demand for that kind of commercial contract that we're seeing now, some being closed, many more in the offing. And so owning gas is a highly commercially attractive opportunity set and it's a necessary one for reliability.
- [04:04]
A
So two assumptions you guys, underlying what you guys are saying is one, so you think that the demand growth really will materialize because all this is based on projections that all these data centers really will get built. And the data centers need power, right? So they don't have the power, maybe they won't get built in the first place. Kind of dependent on each other. And two, Adil, you mentioned that eventually, you know, I don't know what that horizon is, but if it's only medium term, that the supply of generation will eventually catch up. And so at that point we wind up going back into a position like we're in today, we were in two years ago, where the gas assets again are disfavored. What's the kind of longer range outlook? I mean it's one thing, these are assets that people want to recover their investments on. 20, 30, 40 years, what's good for the next 24 months may not be sufficient to make a return on a gas fired deal. Which obviously people think that I'm wrong, that that's not going to be the case. I'm not really taking that, I'm just being a devil's advocate. But just want to justify their investment because there's a lot of very smart people putting money into the business now. So they must think that the long term returns are there.
- [05:20]
C
Well, I guess, look in terms of, I guess our sector, as you know, valve, it works in cycles and it's fair to say we are at right end of the cycle nowadays. Look, having said that, I think what is becoming evident is as gas generation is no longer considered transitionary power source, but I think there is an internalization across the sector that the gas generation will be here for decades to for us to actually transition to the next stage of electrification. So that's, that's also somewhat driven by the geopolitical events and the thinking around affordability. And when you combine all those, I think part of the reason that the valuations are increasing not only because that the market is really great for the next four or five years, but also the outlook has opened substantially and now buyers of thermal assets are not really thinking that I might be the last one who is holding onto this. And they are thinking that okay, in my investment horizon I can make sufficient amount of returns and exits at the end of my horizon. That's important. In terms of demand, there are many different forecasts and even if you actually take the kind of the downside cases, what we are seeing is we are still going to add tremendous amount of demand no matter what. Because if your base is zero, it doesn't matter if it's 3% growth annually versus 5% in the scale of US style country, that's a tremendous amount of generation that's required. And we also think that if there is going to be any curtailment it's probably going to be power driven. So if these demand forecasts will not be realized only because the power sector cannot keep up with actually the data center build out from where we are sitting today, talking with the data centers, hyperscalers, power developers, that's what we are seeing.
- [07:26]
A
What's been the directional change obviously depends on the project and some of it's confidential. But the change in what people think the general returns are on a gas fired power generation facility. I mean so, so generic. If I ask the question, it's probably hard for you to answer anything but let's say in PJM then. Yeah, because that's where a lot of them were built.
- [07:50]
C
That's a fair question.
- [07:52]
B
Just for a second. What I might cut in on Todd is it goes back to what you said about how long do they want to own the plant for or how long do they think they will be it or will they be the last buyer? I go back to when we first formed our firm a little over four years ago and we did a deal for one of our clients that it was a continuation fund vehicle for a portfolio of gas plants in California. Great business, grown really well. But at the time most of the parties we were going to to invest in that fund vehicle or if we were trying to do a sale, they weren't sure gas would be at all in California for four years. We were saying it's 40 years. And so today I Think part of the return is, well now it is, you're not necessarily the last holder. They're viewing it in the form of like the, the length of the asset life which is going to be more like 40 plus years. And that, that was a real step change. Right. And so when you think about returns it also is like well how do I think about that exit and what does it look like? And so it's just a very different dynamic on the overall timeframe and return piece of it before we dive into some of the, some of the numbers and market specific points.
- [08:52]
C
Yes, thanks John. Well, look, in terms of respect to returns, I guess what we are seeing is that you have publicly traded IPPs who are emerging as winners in many private M and auctions. And when we look at the, the private MA transactions where they are trading, they are typically for newer vintage combined cycles. They are trading at around 8.9x EV to EBITDA type multiples. And when you look at the private IPP, sorry public IPP multiples that are trading at 11 plus in some cases 15, 16 multiples, it's very obvious where the arbitrage is right. So the private to public arbitrage is very obvious. And a frenzy today we are seeing is because of that.
- [09:39]
A
Who do you think the buyers are then generally? Do you think it's going to be the publicly traded companies that are going to come in? Because I've seen a lot of private equity funds coming in and buying from other private equity funds who had bought five, seven years ago or 10 years ago and now the normal cycling between them. But what you're saying makes sense to me. If there's enough money in the public markets, that's clearly where it should wind up.
- [10:02]
C
There are three types of buyers emerging in recent auctions and processes. The first ones are, I just mentioned, Those are public IPPs and private to public arbitrage is attractive. The second type is regulated utilities who are real time talking to data centers and feeling the coming demand. And they are basically looking at their supply portfolio and seeing the deficiency and they are trying to bridge that with acquisitions. So that regulated utility universe is also emerging to be another attractive buyer set. And the third one is financial sponsor backed private. Some of them are just getting to a scale where IPO might be considered a suitable exit alternative here. So those three buyers are very topical in today's MA market. We are still seeing financial sponsors active in select processes. But when you think about large processes today and especially baseload generation assets, those three categories are the ones that stands Out.
- [11:13]
B
Yeah. And I would add, you know, on a case by case basis where there is a significant disconnect from maybe where the forward curves show and, or lack of forward curves and. But a very, very strong entry from data centers. For instance, we sold two gas plants recently out west and those two were really best held in the hands of parties that were likely to ultimately contract with the data centers. But in the intermediate term frame they could take that some of that commodity trading risk. And so we also see firms that have that commodity trading capability and are asset owners as also potentially a very viable bidder for assets, especially where the commercial piece may be down the road a bit more in terms of before they contract with utilities or the data centers.
- [11:57]
A
How are the either investors of acquisitions or people looking to develop new gas fired assets? We haven't talked about that yet, but that's also come back roaring back. To the extent people can get turbines, how do they deal with the offtakes? Because when we were doing these deals five, ten years ago, people had all these, as I was saying and when we get in all these hedge, it wasn't that easy to get a ppa but you could get different types of hedge products that were offered. How are people looking at that today given that these data centers, if you've got power anywhere near fiber like they want a data center there.
- [12:34]
C
Yeah, let me, let me start and then I'm sure John will have some ideas as well. But look, you know, in terms of today's market, when combined cycle capital costs at $2,500 per kilowatt to $3,000 per kilowatt, it's tough to get these units built on fully merchant basis in many markets. So at the locations where there is an alignment with hyperscaler data and load requirements, we are seeing 20 year long term contracts at quite attractive return rates for developers emerging. So today's market as opposed to 2010 is actually offering long term PPAs for combined cycle developers. And that's I think the biggest difference and also the second, second biggest difference is the lack of equipment. Where you go back to 2010s you had OEMs racing with each other to offer attractive terms and in some cases equity partnership. Today it's the other way. You need to actually invest in substantial deposits to secure turbine slots. Way ahead of the construction financing. And we are seeing that as emerging bottleneck in combined cycle and gas generation developments.
- [13:59]
B
Yeah, just to take that point a little bit further. Ado completely agree. And so with our clients, our developer clients and sponsor clients in Some cases as well that have these opportunity sets, you really do need to procure that equipment today in order to be able to really go risk off with your project. So you really have to come risk on with equipment that are long lead time items in order to be able to actually fntp your project in the timeframe that the buyers of the power, the data centers ultimately need it in. As often is the case, you know, necessity is the mother of invention. And with our clients, we've really gone out and found private credit lenders that will advance significant amounts of dollars against those equipment purchases with relatively small amounts of equity or no additional equity from our developer clients. And recently these also tend to include not just CCGTs but CTs, RECIPs, batteries. And really part of the story is about behind the meter or at least behind the meter for a while for many of these ultimate offtakers. And so there is a very different approach that, that the, the new world has to take in order to access those PPAs and to get the attention of those.
- [15:14]
A
So we, we hear a lot of stories about the long lead times required to get things like gas turbines and transformers. If somebody has a project that they're looking to do now or somebody's got, you know, hyperscalers interested in a data center, what are you hearing in the market in terms of how soon somebody can actually, if they're just starting today, how soon can they get access to a turbine or a transformer? I mean, how many years out are we looking for you to be able to even start construction?
- [15:45]
C
I mean, look, there are cases where I will talk about the generalities and then there are obviously some exceptions, but we are seeing five to seven years delays in terms of getting an equipment secured. Obviously in some cases where an equipment becomes available because reservation just dropped for some reason or another. But in most of our cases we are seeing deliveries between 2031, 2033 for.
- [16:14]
B
Yeah, I think we see agree with Adle on that. It depends by the parties that you're, the OEMs you're dealing with. Think the typical OEMs that we've usually financed that's expanding, right. And then we're also looking at solutions that are not necessarily new deliveries, but in fact refurbished engines. We've been, we financed a number of those with some of our great clients and some of the companies that produce those that we worked with closely. And I would say as well, we know and are seeing it in a number of our deals, equipment that's overseas now, but will be here once it's kind of retrofitted referred to us standard zero hours, et cetera will be a part of this equation in order to meet the demand. So it is an interesting market right now where the developers can't just assume that they'll have an equipment order that'll fit with their project. They really have to be well out ahead of that and being creative where needed when the OEM timeline doesn't fit theirs.
- [17:13]
A
Or maybe what I've seen also is that they're looking at higher heat rate type equipment that people wouldn't have touched two years ago. People would be very surprised that anybody would have an interest for 800 megawatts of absolutely efficient equipment or backup. But now all of a sudden it's whatever you can get your hands on is good enough. Yeah.
- [17:33]
B
And at the end of the day reliability is much more important to many of these clients than an incremental cost for them. Right. And so some of that older equipment's pretty great in that regard. Right. And so we were that was looking through one of the OEMs, put out a list of what's available and when in terms of delivery. It's a little insight of what of what Adle said but you know what's on a page is and versus what you can actually do is different thing. Yeah. The on that list, you know the earlier deliveries were these but still 28 or 29 were these much less efficient high heat rate engines.
- [18:02]
A
So let's say you have a higher heat rate type product that eventually is going to be less competitive. And some of these data centers also don't have hyperscalers as tenants. You know where does the. Again, you know you're not going to be able to give me a. I'm not asking for a specific answer because you know it's all very dependent on all the facts. But when people are trying to weigh things about what, what's financeable. You know, if I walk into your office and say John, you know can I can I get this deal done? I can, you know start in 2029 instead of 2032 or whatever if I can use these engines as opposed to these ones or I can sign up a contract tomorrow with these guys who are not don't have hyperscalers as tenants. But if I want to wait for the hyperscaler I gotta wait two years. What are the big driving factors there without you need to see it. I'm sure someone has to go talk to you to see the case dependent. But what are the real things that you look for to make sure if a project's going to be financeable and then how does it impact how you would take the deal to market?
- [19:06]
B
Yeah, well I'll take a crack at that first datel and then please, please jump in anytime. So what I would say is, you know, first and foremost we're trying to get that equipment secured through financing that'll really look just to the equipment and so that will give our clients a Runway to get the best deal, hopefully an investment grade deal and a well structured one at that with the PPA and the off takers. And then I think when you, if you don't have that and we've closed deals certainly both in thermal and renewable, without that, then we look at what are the ways in which we can have structural credit support or mitigants in the transaction structure that would effectively allow for it to be treated as investment grade. And if not treated as investment grade, could we still find financing that would be attractive? And with the economics that you can find on these and with the right amount of credit support, whether that's an lc, an LC and additional mark to market measures that might be put in place, et cetera, there are a number of tools both on the ways in which the credit support could be provided by this non investment grade counterparty or parties as well as what can be structured into the financing that may be still a much more value accretive solution to get that project built sooner and lock that in today.
- [20:19]
A
Nothing to add there, Dio.
- [20:21]
C
I mean the only thing maybe to add is we are also seeing increasingly more active and sophisticated private lender space where they are happy to get their hands dirty and understand the credit profile and not necessarily using the conventional metrics that banks or other investors are using. So that universe is very accessible and very hungry to take data center credit exposure.
- [20:52]
A
Does some of that really hinge upon the idea that if this non investment grade tenant doesn't pay, I can terminate their leases and then show since there's a shortage, I can, you know, I'll get somebody else just to take their spot. And as long as that market, as long as I'm confident that the market is good and that there's a shortage data halls out there, you'll find a replacement, right?
- [21:17]
C
I mean look at the time of scarcity for powered land. This is a very liquid asset. We are right now, for example, in the midst of a process where we are selling a portfolio of crypto farms that can be very easily converted to hyperscaler hosting data centers. And we are seeing that this is A very fungible asset class where it can be used by for many different purposes. So it's correct that there is some underlying collateral value there that lenders are taking comfort.
- [21:51]
A
Yeah.
- [21:52]
B
Increasingly hyperscalers that had issued like the, the gas component of the, of the power equation have begun to embrace it at a minimum for transition for many other sites and we've had recent discussions and so it is, you know, is increasingly liquid as Adel said, because the buyers are increasingly ready to use as needed recipes to, to, to, to provide the compute they need.
- [22:19]
A
How much has there been a change in view from hyperscalers that or and utilities generally from the market that there is no room in their portfolio to add more car CO2 to the atmosphere? Basically, you know, you, if you were to talk to people three years ago, two years ago, a lot of the banks, a lot of the PE funds, a lot of the utilities, a lot of the hyperscalers, you know, they were, they were kind of counting carbon emissions and trying very, very hard to only contract with renewables. But clearly with the demand for gas generation that can't be the case anymore to the extent that it was. But how much do you think that's changed and is it a change? I'm not asking permanent because nothing's permanent, but you know, is that a change we think we're going to have here for the next five to seven years because you're saying to get a turbine there it's going to take five, seven years or is this something that you guys see as who knows. And two years from now we might be looking at a renewed focus on counting carbon emissions.
- [23:23]
B
But I'll hit, I'll hit one. You know, one of the hyperscalers that we were speaking with recently, I thought framed it well, which is we still have these long term sustainable objectives around CO2, et cetera. That hasn't changed but we have a business to run and we have compute to ensure that is getting, is getting taken care of. And if we need to do that with other fuels and in different ways than we were originally planning, then we're prepared to do that. And so I think it will be very opportunistic and I'm sure one hyperscaler will be different than another.
- [23:56]
C
Yeah, I think it's really two teams that are interconnected are emerging. The first one is affordability, the second one is growth. Certainly if the further decarbonization is the constraint in front of near term growth, I think there is some reconsideration and we are seeing there the second one Is look, and with the increasing bills, electricity bills on ratepayers, there's a strong push for bring your generation style solutions. And that is also making hyperscalers and data centers think twice and they appreciate that the easiest bring your generation quote unquote could be a peaker or receipt. And that's driven by the policy. And I think that's what's going to happen in the next few years in the long term. I think the trend towards decarbonization is secular and it's there. But I think what we are seeing right now is a recalibration of that.
- [25:01]
A
All right, so I'm going to ask you guys kind of bigger picture because I know you guys are across many parts of the power generation, not just gas, but how many years do you think it's going to take based on what we see here for basically for all the sponsors out there to develop enough energy to put less upward pressure on pricing of energy. Electricity rates, retail electricity rates are wholesale for the data centers. Because to me what I keep waiting to see is the political backlash. You know, they're starting to get some aspects, power prices go up, people are going to be complaining and like you said, the demand is, you know like nothing that we've seen in, you know, in, in my career at least. So it seems like there's gotta be, you know, at some point there's going to be enough pressure on both sides to, to get more of an equilibrium between supply and demand.
- [25:55]
C
Right. And we are seeing some signs of that. Right. So even the PGM capacity auction price cap as earlier signs of what you are just describing. But look, what we are seeing is clearly unprecedented. The analogy is probably transcontinental railroad development or tailgraph development style industrial milestones. So this is something that I anticipate will require in one fashion or another intervention policy intervention. It could be in the form of bring your generation. It could be in the form of a wave of nuclear build outs. There are on the horizon, there are solutions or levers that policymakers can pull. But in the near term, in the next five, 10 years, I think the stress point is us cannot afford to leave behind in this race. It's a geopolitical race. AI is clearly important aspect of economic competition and a global scale. So I think the stress point is that if you dial down data center and AI related demand growth then you might just risk getting left behind. And that's also not ideal. So I think for the near term we are going to see these elevated prices.
- [27:13]
B
Yeah. And the practical answer way to answer it is also just looking at the orders that we're seeing placed and getting placed now. Right. And so we continue with our clients to see the equipment being ordered and we continue to finance that equipment and you know, we continue to finance the build out of these projects. So it would appear that they will be built at least, you know, to 2029, 2030.
- [27:36]
A
Right.
- [27:36]
B
Based on likely the next one to two years of at a minimum of FNTPs. And if the trillions of dollars that are set for this space, a small portion of that of which will go to build power plants, but still large number by power plant standards, this appears to be something that should be a at least a five to seven year cycle because of the timeframe to build. Now as you get to recips or CTs, that cycle is obviously a bit smaller. But with the delay in delivery times and even with the used, refurbished and other equipment that's kind of going to fill the gaps. It appears to be a multi year cycle here that we're entering into of
- [28:13]
A
significant build a positive note for most people in the industry and means that we got a lot of work to do. So thanks for joining us today guys.
- [28:21]
C
Thanks for having us.
- [28:22]
B
Thanks John. This is great.
- [28:25]
A
You can find us online at www.projectfinance law or send us an email at currentsordonrosefulbright.com Please rate, review and subscribe on Apple Podcasts, Spotify or your preferred podcast app. Our show today was produced by Emily Rogers.
- [28:42]
B
Stay ahead of the current.
- [28:47]
A
Sam.