← 返回

2026.09.24 | TB | Power | NRG Data Center Large-Load Deals

原标题：NRG Energy – Data Centre Project Tailwinds & Large-load Contracting Dynamics

#power #datacenter

Third Bridge · 2026-09-24

Analyst-led • 16 Sep 2026

Companies: NRG US, VST US, CEG US

Mark Herrmann (MH) Head, Trading & Commercial Operations, Americas at Jera Co Inc

Anthony DeRuijter (AD) — Third Bridge Sector Analyst

Agenda

Key Insights

Introduction

AD [00:00:02]: Let me start today's Third Bridge Interview entitled NRG Energy – Data Centre Project Tailwinds & Large-load Contracting Dynamics by introducing myself as the moderator. My name Anthony DeRuijter. I'm very pleased to be joined today by Mr Mark Herrmann, former VP of Commercial Operations, Trading and Structuring at NRG Energy. Mark, thanks so much for joining me today.

Before we begin today's Interview, though, could you please state specifically, I agree or I disagree to the following statement, and the statement is: You understand the definition of material nonpublic information and agree not to disclose any such information, or any other information which is confidential, during this Interview.

MH [00:00:35]: I agree.

AD [00:00:36]: Thank you very much, Mark. Would you mind briefly introducing yourself?

MH [00:00:42]: Mark Herrmann, and I've spent, I guess, most of my career working for about three different IPPs. Most notably, I guess, for this call, is I spent a bulk of almost 14 years at NRG, where I was responsible for commercial operations, which included trading, structuring and origination around the wholesale and retail portfolio.

Questions

[00:01:14] What is your load growth outlook for NRG Energy and other IPPs [independent power producers], focusing on key markets such as ERCOT [Electric Reliability Council of Texas] and PJM [Pennsylvania-New Jersey-Maryland Interconnection]? Could you also discuss the RBP [Reliability Backstop Procurement] in PJM and Batch Zero in ERCOT?

MH [00:01:36]: I think the load growth that you're seeing, obviously, it's a key story. There's a lot of debate around it, but it is absolutely real. What you're seeing for the queue around data centres, expected large load, that's not necessarily the real load. I think the question is really going to be how much of that demand is going to be bankable, how much of that is going to actually be able to be interconnected and on what timeline are we talking about. I think that there's a big difference between the headline numbers that you're seeing out there vs what I would call bankable load. PJM, you talked a little bit about that. That's probably the cleanest of the stories. PJM has got, I want to say, a little over 3.5% peak summer load growth over the next decade and over 4%-plus annual energy growth. At the same time, you were talking about the reliability auction, the capacity auction, the '28-'29 capacity auction came up about a little under 7GW short of its reliability requirement. They have initiated a reliability backstop procurement auction to help fill that gap that's needed for new supply to meet that pretty significant load requirement that's out there that's driven on the back of not only the data centre demand, but a pretty significant amount of demand for just the onshoring of new large load and industrial.

That's what I would call the PJM story. Moving then to the other key market, which is ERCOT, the numbers happen to be much bigger, but there's also a lot more, I guess, I would call noise around those numbers. Some of the preliminary forecasts around ERCOT that you see around what's in queue for new load I want to say out to 2032 timeframe was in the high 300s, almost like 400GW of load expected. Obviously, again, a lot of that's a headline number, and there's certainly some duplication of permitting requests that are in some of those load numbers. I think ERCOT itself is trying to figure out exactly how much of that load is real and realistic. There is a process in ERCOT right now, which is called the Batch Zero process, which is a process by which ERCOT is looking to batch new load interconnection and begin to study them as part of a batch vs an independent discrete study project by project. They're putting them into batches to try to speed up the ability to interconnect new load.

Just recently, there was an announcement that ERCOT has… They announced this to the PUC, they have approved about 66GW of new load to come in as part of that Batch Zero approved process. Just to put numbers around that, you can assume within the next 3-5 years of that 60GW-plus of new approved load through the Batch Zero on top of a peak load that was just reached this year of around 90GW, you're looking at 150GW type peak load market in ERCOT over the next 3-5 years. That gives you a sense of the difference between what's a headline number for load vs what's an approved number across the PJM and the ERCOT markets and then just a sense on upscale on just how big that is expected to be over the next few years, just given what we can see now in terms of current approvals.

[00:06:23] Peak load records were hit in the summer. Given Batch Zero and all the interconnection requests in ERCOT, you'd assume there would be more of a reaction when we're talking about forward curves, but it seems that the reaction has been fairly muted despite the load growth expectations. Where do you think ERCOT prices are at right now, and why might forward curves not be reacting as much as expected?

MH [00:07:06]: I'm not sure it's super inconsistent with the load story because I think the market is discounting that headline queue that people were talking about, the 300GW-plus of incremental load. Now there being some ambiguity around what's really going to get approved, what's actually going to get considered as part of the batch process for getting the new large load online. Then I would also say, second, that a lot of that large load is going to be bringing its own power, so a 1GW data centre may come with a gig, 1.2, 1.5GW of new generation that's behind the meter, serving its own load, and then it may even be upsized to provide some additional generation to grid. I think that people are looking at that as well and thinking through that. Then I would say maybe the third item is that ERCOT is also adding a lot of solar, a lot of batteries. There's new gas capacity that's coming on, so I would say the market is probably pricing both sides of the equation right now.

I think that certainly makes sense. Then maybe another leg that's not a generation leg, but the transmission story, which ERCOT has a process right now to approve new 765kV lines, large lines across some of the major cities in Texas and be able to provide a much bigger backbone for providing distribution of power, which should be a reliability benefit and then that benefit to the system. I think out of the five lines that were proposed for the 765, I want to say three of those five have actually gotten approval. I also think if you look at NRG's own presentation, they were talking about the fact that they don't think that the forward curves reflect the demand outlook. I think that was back in their last presentation where they showed maybe low to mid-50s in terms of pricing for power out in the forward curve and expecting that they didn't see that as being really reflective of the expected forward demand. I would say there are certainly reasons why people are fading that expected impact on prices for the items that I just outlined.

[00:10:18] When we think about the significance of forward curves in pricing structured contracts, is there a need for IPPs, such as NRG, to hold off or wait for forward curves to reflect their expectations of what real demand might look like before going into broader long-term contracting? How do we think about their need for a heightened benchmark by which to price contracts for the existing generation?

MH [00:10:51]: Yes. I mean I think, look, the first thing I would say is it's not like you're just going to see a change in the prices overnight and that you're going to see a sharp uptick in forward. I think that what's happening right now and what you'll see is that reliability is becoming more valuable. I think that value in the market is going to show up, first, in shape, the shape of the power price, in terms of the peak vs off-peak, it's going to show up into basis pricing, like which areas are more congested vs less congested. It will show up in scarcity optionality and probably the way that people start to price bilateral contracts and not necessarily just in a step-stair move up in the ERCOT flat, the strip and the forward prices that you might see on ICE. I just wanted to make that point out there first as it relates to how you typically see prices. It's shape basis, it's scarcity optionality and then you start to see it reflected in new bilateral deals that are done for offtaking for those that have an asset presence, either for load deals or for generator deals. Then just going back to your other question, you wanted to know the how this is impacting what pricing for new generation, or what was your…

AD [00:12:40]: Understanding how the forward curve and market prices are going to be setting the benchmark for reference prices and contracting, if the IPPs don't wait or if they price off today's forward curves, they get a worse deal than if they wait for forwards to reflect their expectation of reality, if you will, with the load growth. Is that accurate?

MH [00:13:17]: Look, I think there's a difference in how folks are looking to contract and maybe this gets into your BYOP. I don't think that you're looking and you're seeing folks, especially on this BYOP and other contracts, you're not just seeing folks saying, "Hey, we're just going to sell energy at the current forward prices. I think that, for example, this BYOP, there is a structure around the contract where there's the offtake is structured in a couple of components. One is a capacity-based charge that recovers the capital and return for an asset. I think in the case of what NRG was talking about, I think a very significant portion of how they think about offtaking is actually recovered in that capacity-like charge. Then the second is that any gas operating costs would be part of a separate payment. Between both of those, then you're getting a much more contracted infrastructure type offtake as opposed to just looking to sell outright the forward curve and hoping that you can get a mark for a 10-year infrastructure offtake.

AD [00:14:53]: To summarise what you said, it's looking like the cost of service model for utilities, in terms of capital recovery and the variables covered in pass-throughs. Is that correct?

MH [00:15:15]: Yes. I mean I think that's one way to think of it. Yes, for sure. It's more of a contracted infrastructure type offtaking arrangement as opposed to just what you saw in the past, which was just more of a merchant player or people looking to hedge off the forward curve for the liquid 3-4 years out and then just taking a merchant bet in the back, like you're starting to see more… You're taking away a lot of that commodity price risk out of these projects and swapping that risk more around what I would consider to be execution risk around these projects as opposed to commodity price risk by being able to structure these offtakes that have a capacity charge component as well as an operating cost pass-through.

[00:16:18] Do you see any risk of NRG being overlevered to large loads moving forwards, and why? Management previously stated that even without the data centres and assuming flat power prices, the company would expect double-digit returns.

MH [00:16:41]: Yes. I wouldn't characterise the company as over-levered to large load yet. I guess maybe to your point, maybe this is where you're going is I think in looking at some of their publicly stated comments, they're saying that the existing business supports a 14%-plus EPS CAGR over the next few years, and that's pre data centre opportunities. I think that they would consider… At least in what I've seen on some of their public statements like the data centre would be upside to that and the offtake of the data centre. I think of things like what they've done with really building out the acquisition model on buying out LS Power, the Rockland assets. I think those types of getting more into the generation side, I think, gives them more monetisation channels. Just taking, for example, the LS Power, they've got a significantly larger platform now to clear that latest '28-'29 capacity auction at the cap of USD 325 a megawatt day. That's a pretty sizeable monetisation channel regardless of any kind of load growth. That's just monetising into the capacity market. Look, they have an entire fleet now between LS and Rockland of additional assets that they can use to support retail, they can use to wholesale.

They can bid into the capacity. They can use it for C&I demand, and then that's all whether or not a certain hyperscaler project actually happens. I'd be more concerned about them being overlevered on the large load growth if they started to look to underwrite a bunch of acquisitions or committing capital around large load forecasts before they materialise and making some of the old school mistakes back in the old merchant generation days and not being disciplined. I think like making sure that they're looking to build when they have contracted load or contracted offtakes and being disciplined around that is the right strategy to go. I haven't seen that they've done anything outside of that.

[00:19:38] How much bigger has NRG's generation platform gotten with its recent acquisitions? Are there any risks to point out in regard to previous build-bust cycles that might apply to the company today?

MH [00:19:56]: Yes. I mean I think the big risk would be what I mentioned, which is if they did start to go out and acquire without having an offtake or a strategy or they started to commit capital around new build without actually having that contractability, I think that would be certainly a risk. Otherwise, yes, I don't think so. I mean I think it fits in really well. Those acquisitions fit in really well, and alongside with their CPower acquisition, too, which gives them more controllability around load. Now you've got more generation in some of the key core markets in Texas and in PJM. Then you've got the CPower, which allows you some controllability around load to manage that integrated gentailer model. All that certainly makes sense from a strategic standpoint.

[00:20:56] If there were issues around the data centres potentially not materialising as expected, or the offtakes perhaps not coming at the scale needed to fully contract all the new capacity, what would be the role of NRG's retail platform? Would that be enough to soak up all the new capacity? How should we think about the scalability of that as a backup?

MH [00:21:27]: Like I mentioned, having this generation and having CPower, it gives them a lot more optionality in their portfolio and regardless of whether they get the hyperscaler offtake. I think the strategic benefit isn't for them just to be longer power, I think that they get a much bigger physical portfolio against which they can sell more retail load, they can sell more C&I load. They can certainly provide data centre load products and offerings to data centres. They can provide the capacity products as well, like we're mentioning either it's selling through capacity and bidding it into the capacity auctions or the third-party capacity products. They can provide structured transactions and make markets that way bilaterally with other counterparties. I think that having all this, having that optionality, yes, it gives you hyperscaler opportunity, but it also gives you those other levers that I mentioned as well. I think that the acquisitions that they've made on the LS Power and the Rockland, it also helps in terms of building out the eastern part of their portfolio and then also in Texas as well. It gives them some diversification there as well in terms of location.

[00:23:17] What are you currently seeing in regard to data centre contracting structures? What is the average duration of a contract? How are the rates being structured – essentially the capacity and energy components you talked about – whether it's fixed price or indexed to something? Do you see any premium discounts, or is it level to a benchmark or reference price?

MH [00:23:52]: Yes, there's a lot there. I mean, maybe I'll try to break it down into different pieces and then we can keep walking through that. I think you talked about the different types of contracting. I think there's a couple of different contracting markets that I see developing for data centres. I'll start there. I mean the first is the existing clean generation, which is mainly nuclear, like you're seeing a lot of very traditional long-term PPAs. You've got Constellation's Clinton agreement with Meta. I think that's long term. That's around 20 years for over 1GW. You've got Vistra's Meta agreement that's, I think, 20 years for, I don't know, like 2.5GW of existing nuclear. Then you've got Talen's AWS offtake that runs through, I think I want to say, beyond 2040. Those are more PPA like. They're longer term and they're for existing nuclear, that's one offtake structure I've seen on the data centre. Then the second is really around new build thermal, which looks a lot less like just these traditional fixed price PPAs like we talked about that you're seeing on the nuclear.

They're really more of infrastructure type contract or a toll type contract. In those you may have a fixed capacity payment like we talked about that have a fixed availability and a capacity that covers a certain amount of expected return. Then you have separately operating expenses that come through as a separate variable payment for both fuel and OPEX. I think those are very different type of offtake structures than a traditional fixed price PPA. On top of that, you may see other pieces that include maybe an extension or an expansion right that's on top of those deals, you might see a minimum payment obligation, some sort of like a floor on the expected payout. There's oftentimes different development milestones or incentives to go for different milestones. There's credit support, there's termination payments. There's backup bridge power, different arrangements and one-offs that can be included for those needs as well. Then any additional contracting on top of the thermal that's for needs around renewable or any clean energy attributes, those are also structured in alongside with any BESS and energy storage that's needed as well.

Those are some different things to think about and the differences that I've seen across various data centre offtake structures and how they differ from the clean nuclear vs the thermal and then the structuring and then maybe some of the other one-off expansion or extension options.

[00:27:43] How should we think about the typical data centre contract durations we're seeing now? NRG's BYOP [Bring Your Own Power] project is a minimum 15-year contract. Would you say that's where we're generally seeing average contract lengths?

MH [00:28:09]: Yes. I mean definitely, these contracts are longer in nature, 10-plus years for sure, on the new build thermal. That's what's needed and necessary for financeability. Then what you see on top of these, though, is they'll often structure additional option rights to be able to expand out and have a right of first refusal for ability to expand out for another five, 10-plus years at a certain price or better.

AD [00:28:49]: That would essentially be a new greenfield asset. I would be curious, and this might change how the contract is priced because it's perhaps more difficult to model out the capacity and OPEX or energy components. However, what if you have an existing asset? Would it follow similar structures as what you called out earlier, or are those contracts going to be drawn up and priced differently?

MH [00:29:12]: There's all sorts of different way. If you're talking about just, hey, how are people able to offtake an existing asset, whether it be for a hyperscaler or just as a hedge on an existing asset, there's many ways to contract those. You can either fully toll those assets out and either give the dispatch rights, hand them over or financially toll them. It could be either physical or financial. You could sell them out as a heat rate call option and sell the option against the asset. You can do a structured offtake, where you can set up a structured deal for a certain amount of years where you can offtake the asset either just to the energy only or with capacity. Again, with dispatch rights as a physical toll or financial, which is just financially settled. I mean there's many, many different ways to offtake it. I think what you're seeing though is a move, especially with NRG and others where there is more of an interest in being able to offtake longer-term, longer tenure deals with some fixed component, capacity type component to it and some sort of a fuel and OPEX pass-through.

There's many different ways that you can decide to offtake, or you can decide to just sell straight up power against it and buy the fuel against it and then you can decide to do that in a much shorter-term basis and through standard market products that are tradable. It's anywhere from tradable to structured products to fully contracted infrastructure, like offtakes. There's a gamut of opportunity.

[00:31:37] Are there different revenue levers embedded into a hyperscaler or data centre offtake for an IPP? Are capacity and energy going to be established upfront and the cash flows are visible in that sense, outside of some variability in OPEX? Is there no other way to earn more on a contract?

MH [00:32:07]: There are so many different ways that you can structure them. I mean, it would be tough to say, but I would say if you're asking about what's… Are you asking specifically about what's the earn-out capability or how are you seeing the market structure offtaking for power that's used to serve data centre?

AD [00:32:32]: What are the additional ways to earn more on a contract with a data centre outside of how it's priced on capacity and OPEX? Can you add storage or some sort of additional redundancy for additional reliability or flexibility and you can price for that, or is that already built in from the get-go?

MH [00:33:03]: I mean, to be fair, I don't think that… They're not really looking to do the offtake as a way to earn, if you will. The way they're thinking of it is, I need the power to be reliable and to be on time, and it's all about speed to power to actually get these data centres up and running and powered. I would say, yes, they're looking at different things they need to do to manage the structure so that it meets the reliability needs, it meets the speed to power needs and it meets the lowest total cost of power. As part of that, yes, you're right, like they are including renewables, they're including BESS as part of it to figure out what is the optimal lowest cost and most reliable way to provide power to the data centre. They're not so much looking at ways to structure things so that there's an earn-out if they can make more money on the power or whatnot. It's more about getting reliable, low cost and speed to market on the power. Does that help? In some cases, you're seeing data centres that are willing to say, look, we'll go behind the meter, and we'll do a five-year payback on the new power that's needed to get built just because they're so wedded to making this work in a short period of time that they're willing to structure it like that.

AD [00:34:54]: On the more favourable end of data centre contracting, you would imagine there are some data centre customers that would enter into an offtake that's priced for a five-year payback for the IPP, but I'm assuming the duration of the contract is still going to be on the longer side, such as 10-plus years. Is that correct?

MH [00:35:13]: Yes. Yes, exactly. Yes. The contract is longer, but they are accelerating the payback of the generation asset.

[00:35:25] NRG highlighted that its first BYOP data centre project in Texas has advanced into Batch Zero. What are your thoughts on the contract terms? How favourable do you think they are vs what you've seen more broadly and have highlighted? Some of the key terms would be the minimum 15-year contract length, with options for extension and expansion, and it's priced for a 12-15% return. Overall, it's a big project – over USD 3bn – and I believe COD [commercial operation date] is due for late 2029. With the project parameters and contract dynamics in mind, is this favourable net-net? Is it in line?

MH [00:36:23]: I think the way that they structure it is very important relative to the headline that they have out there. I agree. I think it's a sizeable investment, but what I like about the way they structure it is that it's not just relying on ERCOT spark spreads and the forwards to earn that return. They've done what we talked about, which is they've structured it more as a contracted infrastructure type investment vs the old-school merchant generation play. The customer pays that availability-based capacity charge. That capacity charge, I think they've said returns around 95% of the project's required returns. Then on top of that, they have a gas and operating cost that sit as a separate payment, and all this backed by an investment-grade parent guarantee. I think that's a much more stable type investment. What they've done is really taken out a lot of the commodity risk and transform the remaining risk really to be execution vs commodity.

Execution risks around document, getting the permits, gas infrastructure, construction costs, equipment performance, interconnection and keeping the overall project on schedule as opposed to just being concerned about the commodity and the offtaking side. That's my understanding of how they've structured that contract. I think it makes sense as an infrastructure play and it delinks a lot of the commodity risk.

[00:38:38] My understanding of NRG's BYOP data centre project is that it's a single undisclosed hyperscaler, and the company noticeably talked about its credit rating, which brings up counterparty, credit and concentration risk in a single customer. How do you see NRG managing that? How would most IPPs manage that type of risk?

MH [00:39:06]: Yes. Look, I mean, there's no doubt just having one hyperscaler, I mean, understanding the credit risk on that is key just given the size of these projects and ensuring that there is the right parent guarantee with the right entity and especially given you're talking 10-, 15-plus year offtake contracts. All of that is is clearly like a big component of really analysing and understanding the overall economics and risk, so no doubt about that. Look, as valuable as the 15-year contract is and how you structure the contract being this mix of a capacity type construct with the pass-through for all the operating and fuel expense, like as valuable as that is, just as valuable as making sure that you understand the credit standing behind the actual payment obligation, especially considering the length, tenor and size. Yes, I mean, that's definitely critical.

[00:40:38] In NRG's Q2 2026 earnings call, the BYOP project is largely priced for return thresholds, but management commented that the rough PPA [power purchase agreement] equivalent for contract pricing on this project would be in the USD 85-90 per megawatt hour range. If you were to see that pricing level on a 15-year contract, how should we think about not just how favourable that is compared to a broad-based PPA pricing and PPA terms in ERCOT, but how feasible that would be?

MH [00:41:24]: I think a lot of people look at that and say, "Oh, look at that relative to where the current forward markets are. It's a massive premium to where the forwards are, and obviously, the ERCOT forwards are materially lower than that. What you're talking about, and I want to say, I don't know, just looking up… Your forwards are in the high 50s, whereas you're comparing that to an USD 85-90. What you're talking about here is a very different product than just a forward energy product. A customer that's buying this BYOP is buying something different than the forward, so you're getting dedicated new generation, you're getting availability, reliability, you're getting a long-term, I guess, I'd call it reservation of capacity. You're assuming that you're also managing the construction execution and getting that path to power and that speed to power that I was talking about. That's what you're buying when you're paying the USD 85-90. Really, I think a big part of that is like what is the ability then and how conceivable is it to get a new build at today's construction costs and does the USD 85-90 cover you given those latest construction costs.

That, along with all the certainty, the reliability, the execution, that's all of what goes into that USD 85-90. That's your real comparison for what the hyperscalers is buying vs the mid- to high 50s ERCOT north forward market. That's the difference between what a reliable gigawatt that's going to be available to me at my data centre for new power for the next like 10-15 years, like what's that going to cost me vs what's my ability to buy power off the forward market. That's the difference between what you're talking about product vs product.

[00:44:14] The BYOP project's contracting model looks quite similar to the cost of service model we see with investor-owned utilities in the US and more broadly. NRG stated that almost all of BYOP's projected free cash flow would be supported by capacity payments, rather than utilisation. Utilities earn roughly, let's call it, 8-10% nationwide. The company is targeting something way higher than that. What's the feasibility that NRG can offer or benefit from a similar type of revenue model, but earn a much higher return? Where is the risk that the company takes if a lot of that is just priced out in the contract?

MH [00:45:16]: I mean, if the question is like, hey, why is NRG… Why are they able to get a higher return than a utility? I mean I would say that for this type of a structure, it's that, hey, they are a competitive player, and they're being asked to build and take on all of those risks that I talked about previously in terms of construction, execution, all the costs, the cost to build, taking on the risk for that and taking on all those risks in a market that itself is highly risky and doing it on the back of a sub-investment-grade company, I think NRG is still sub-investment grade. They would then require a higher return for those projects than a utility. Look what's happening in these markets, you're hearing the comments, I think it was from Paul Segal from LS Power, talking to PJM about the fact that all of this reliability backstop, all this ambiguity around the capacity markets, which were meant to provide certainty for investment in these markets have not done the best job of providing that certainty given that you've traded at the cap three auctions in a row and haven't been able to incentivise enough new generation to come in. What does that mean?

That means that your risk capital, your cost of capital is going up, and it makes things more expensive for those that are going to take on those risks. For a competitive player, that's very different than for a utility that has a rate base that can put out… They can build and get a regulated return on their investment for whatever they build. I think that's why you're seeing the competitive players like the NRGs of the world require a higher return. Frankly, like I think on the flip side, the folks like the data centre players and the hyperscalers, I mean, they're willing to pay a premium because of the value of power and the fact that it's critical for their ability to provide compute, no power, no compute.

[00:48:00] What are your thoughts on NRG's positioning vs the other major IPPs, especially for data centres or any other elevated load growth? Looking at the big three framework that NRG is in with Constellation Energy and Vistra Corp, those players have nuclear and NRG doesn't. Would you say that's a disadvantage for the company?

MH [00:48:26]: I would say NRG's advantage is very different from Consto [Constellation Energy] or Vistra. I mean, yes, like you said, both of those platforms, Constellation and Vistra have really valuable and large nuclear fleets. NRG did have a nuclear unit, STP, but they've since sold it. Look, I think Consto [Constellation Energy], those large nuclear fleets for Consto [Constellation Energy] and Vistra allow them those PPA-like offtakes that are long duration and they're carbon-free. Then I think their 20-plus year PPA-like offtakes demonstrate that value. Those have all been in market and people have seen what those have generated in terms of offtake value. I think NRG's platform is different, but they offer a more full commercial platform. I would say what they offer is, they've got a huge retail book. I think now with all these acquisitions, they've got, what, back to 25GW-plus of generation. They've got the CPower demand response component as part of their platform. They've got a big gas fleet. They've got ability to do development, and they've got access to new turbines and EPC capacity for being able to build out new generation.

I think about that, and I think like they definitely have a nimble diversified portfolio of opportunities that maybe some of these others don't have, so it's just a different book and a different offering. The latest I read is NRG saying they've got little under 5.5Gbit of new build development capacity that's with a line of sight on turbines and EPC that goes through 2030s. That's an interesting advantage for them that's very different than what you're seeing from the others I mentioned. I think it's, hey, if a customer needs 1GW of zero carbon existing power, then clearly, they're going to be calling Constellation or Vistra, and they have an advantage. If the question or opportunity is, hey, I need 1GW of power, but I need it at this site, I need it… I need bridge power. I need to have a suite of opportunities, that might include some renewable, it might include some thermal.

I need someone that has interconnection access, and I need somebody that can also operate those units for the next 20 years, then I think NRG is one of the offtakers that you're going to contact, and they have a very different and credible offering. Look, do I think that them not having nuclear is a disadvantage? I think that was like the second part of your question. I don't know that I would call it necessarily a disadvantage. I would just say that they have a very different offering. I think some of the earlier deals of those existing nuclear, I think they've been struck. I think the next phase of the market is going to be looking for solutions that are going to be more localised and probably more bespoke. I think having a more portfolio of those different flexible options, I think it does give NRG some advantage.

[00:52:53] You mentioned speed-to-power being an important consideration and demand of data centres. It seems there's a lot of emphasis on BYOP when it comes to IPPs and utilities – traditional grid-scale generators serving data centres. If that's the case, that probably runs counter to speed-to-power, so is there a risk that alternative distributed energy solutions, people with reciprocals or repurposed aeroderivatives, modular simple cycles such as Solaris or even fuel cells, such as Bloom, become more of a competitive risk to NRG and eat into its opportunity set?

MH [00:53:41]: Yes. I think folks that are doing fuel cells, folks that are doing renewables, my understanding is a lot of the data centres are going to need a mix of power. Yes, gas generation is going to be a big part of it, and it's going to be a backbone for having the dispatchability that's necessary and the reliability that's needed, you're going to need a gas component. While having a low carbon component isn't the… It's kind of moved from being the most important to being… It's still important, but it's not probably the leading indicator, that's still a part of the overall power solution, the need for low to no carbon sources in the mix. I think it's going to be gas in combination with some renewable and some storage at a lot of these facilities. I think NRG being able to work with those different entities as a partnered solution provider, I think that's probably where I would see things migrating because, look, they can also benefit, they being the battery storage providers, the renewables, the fuel cells, they can benefit from the fact that NRG has the land and the interconnection rights on their sites.

If they can leverage that, then there is a partnership play. I think there's opportunity, I think, enough in there for everyone. I'm sure that NRG is trying to optimise the fact that they have the optionality at the site itself.

[00:55:41] Are you anticipating any potential indirect competition, where if data centres and their need for speed-to-power incentivises the proliferation of alternative or behind-the-meter generation equipment – whether it's reciprocals, gas-fired equipment or other types of more modular solutions behind the meter – does that potentially sit on the market longer term, even if it's just a bridge and the data centres ultimately decide to connect to the grid or go with grid-scale power? Could that be a supply risk at all?

MH [00:56:38]: Is the idea like, hey, how much can these other recips or fuel cells eat into NRG's opportunity. I think if you're talking about some of these large campuses, I think NRG is, I'm sure, going to be looking to bring recips as well as CCGT. I don't know that it's going to be a CCGT vs recips vs fuel cells type market. I think it's going to be probably, let's get the modular power in first because it's fastest. Then I think that there's going to be a need to complement that with the higher efficiency generation, such as the CCGTs. I think it's going to be a function of like who's the right IPP to wrap all those pieces together and that has an advantage. I think that that's probably the route like an NRG or their competitors are going to be looking at is that, hey, we can wrap this together. We can bring what's needed first. We can also bring the high efficient, and oh, by the way, we have the land, we have the interconnection, and that brings the locational advantage to us and the speed to power advantage to us as well. It makes it real and makes it bankable.

[00:58:17] In the long term, do you see any potential risk of price competition with low-cost generators, such as renewables plus storage, developers, players or IPPs that primarily lean that way, especially if storage durations extend?

MH [00:58:46]: Yes. I mean, like I mentioned, I think it's going to be a combination of renewables with BESS with gas long term and some fuel cells that's going to be needed to serve the massive demand. I don't think it's going to be a one or the other. Again, I think it's going to be, yes, you bring the modular first, you bring what's on fastest, and then obviously, you supplement with a more efficient generation as you go. Looking at the renewable story, yes, it's lower cost. It's carbon-free, but there's a difference between the value of energy that comes from dependable capacity, and dispatchable capacity. I think that you're going to need a portfolio that is less about just maximising the megawatt hours, and it's going to be more about maximising the availability, the reliability, the firming and the capacity optionality that's needed for a site. Certainly, it will include renewables. I don't think it means that renewables are going to eat the lunch, if you will, and the margin out of these other dispatchable resources.

I think it's going to be a combination of all, and I think there's a shift, like I said, into valuing what it's worth to have dependable capacity. Look, I think that's why you're seeing this BYOP type structure with NRG, where they're saying, look, we want to get paid for the megawatts that we have available and get a capacity payment that recovers for that availability, as opposed to just whatever the economics are for the by megawatt hour of what we produce. They're saying, you're going to pay us for stability, for capacity, for reliability and for true speed to power and access to real megawatts that are dependable and dispatchable. I think that helps them to mitigate some of this risk around renewables and other forms coming in and potentially disrupting the project economics.

AD [01:01:31]: We will now conclude the Interview. Thank you for joining me and sharing your thoughts on NRG.

Transcription ends at 01:01:56 of the recorded material