All righty. Hi there, everyone. Welcome to Talen Energy's Investor Day. We're so glad that all of you can make it. Wow, what a crowd! And a big hello to everyone who is listening in virtually in the webcast as well. I'm Ellen Liu, Head of Investor Relations and the MC of the event today. And to my right are the stars of the show, so CEO Mac McFarland, our CFO Terry Nutt, Chief Commercial Officer Chris Morice , and our Head of Cumulus Growth Cole Muller. After management's remarks today, all attendees will have the opportunity to ask questions, whether you're in person or virtual, and that will be moderated by yours truly. But first, a safe harbor statement.
So we've posted a link to this webcast and the presentation on our website, TalenEnergy.com, and we are making some forward-looking statements that are based on current expectations and assumptions. Actual results could differ due to risk factors and other considerations described in our disclosures, and we've provided information reconciling our non-GAAP financial measures to the most directly comparable GAAP measures in the appendix of our presentation. So thanks again for joining, and with that, I will turn the event over to Mac McFarland.
Great. Welcome, everyone. Fantastic. Appreciate everybody coming out, and I think we had over 110 attendees for being in person and more than 100 on the phone, so it's quite a turnout. Appreciate everybody's interest in Talen. You know, as we were getting ready for this yesterday, and we're doing a little bit of run of show, running through this, it reminded me that I've been in this room before. First time was about 10 years ago with Terry during the EFH bankruptcy. We were getting, I think, interrogated by the U.S. Trustee and the UCC, right? Is that right? And then over there, and I'm looking at you, Phil. Over there, I spent about 2 hours getting yelled at by the creditors of GenOn, while as I sat there stoically. You were one of them.
That's okay, we're still friends. But, so we've got a, I guess, either warm, you know, memory of this place or scar tissue. So, later, if you want to make us feel really comfortable, you can yell your questions at us, okay? It'll be like being in distress again. But what a... Trying to walk around a little bit because this room is so big. What a difference 15, 16 months makes, and it's. I'm glad to be back in a regular way company here at Talen. It's been 15, 16 months since we exited, and I think there's been a lot accomplished over that timeframe, and, it's, it's nice to be part of an ongoing operating company as opposed to restructuring. So with that... Okay. IT glitch, that's a good start. So why Talen?
It's the first slide in the deck, if you want to look at it. It should be memorized, so we are an IPP that's focused on being an IPP. We're a long IPP, principally in PJM, which is where most of our operations are, okay? We like PJM because of the increasing demand, declining reserve margins, therefore leading to increasing capacity and energy prices, okay? Chris is going to talk about PJM in the market and our position as a long generator in that market. All right? Terry is going to talk you through how the increasing capacity and energy prices are manifesting themselves in our forward look, both in 2025 and 2026, and where we think things might go from there. So an IPP focused on being an IPP.
Second, our AWS deal, I think, was a, I wouldn't call it a watershed transaction, but it was an inflection point in the industry, okay? It signaled the coming of the AI economy and the need to power it, and I think our first mover position, signing that deal, gives us a strategic advantage. It gives us intellectual property that was developed over years and months of contracting and things that Cole's going to take you through, so with respect to the AWS contract, we're doing two things. The first is, and many of you have heard me say this, we're looking to perfect and accelerate that first 960 . When we look at how people have risk-adjusted what they think is in our equity or in our value, most people risk-adjust that full 960 campus down, okay, so we're looking to perfect and accelerate that.
Cole's going to talk about the ISA, okay, and how we see things and the outcome of that. But we're also excited about the opportunity to leverage what we learned in the intellectual property, if you will, gained during that contracting experience with AWS to other sites and how we might expand upon that. Disciplined track record of safe, reliable, and cost-effective operations. To me, this is table stakes as an IPP. You have to do things well. We operate big, heavy, rotating machines, okay, that are energized, pressure, steam, electrified. This is table stakes for us. We do it 24/7 , 365 . It's what our 1,900 employees do. Now, oftentimes, we come to investor meetings like this, we sort of skip over this because everybody wants to know about the transactions. They want to know what's going on at FERC.
But without our employees, none of this is possible, and so this is table stakes. We have to do that to deliver the financial outcomes that come through. And then lastly, focused on unlocking value and returning capital to shareholders. So many, obviously, you saw the 8-K and press release this morning. We're re-upping the share repurchase program to buy $1.25 billion of dry powder from here. And we can do that on the strength of our cash flows that Terry's gonna walk you through. But we're doing that on the heels of $900 million plus of repurchases that have already been done. So if you look at that and think about that, when you put the two together, we will have nearly repurchased all of the equity that we had upon emergence 16 months ago.
Finally, with respect to that in our capital allocation program, Terry's gonna take you through how we see that going forward from our cash flows. Where's the clicker? Okay, we got a human doing it now, as opposed to me. Our focus, you heard me say this, our focus as an IPP is to deliver the highest value for every MW we have and every MWh we generate. And we think the financial metric that people should look at, in our opinion, is free cash flow per share and free cash flow per share growth over time frame. Okay, I've already heard some comments about this slide, about why we don't incorporate our repurchases in this, but let me just give you three numbers that I'd like you to think about as you look at this slide: three, 10, and three, okay?
First off, over the next two years, we're going to increase our Free Cash Flow threefold. Free Cash Flow per share, threefold. When you get to 2026, that has a 10% Free Cash Flow yield, if you look at the midpoint here. That's $14 per share over a $140 share price. I don't know exactly where we are right now, but let's just go with it, right? 10% Free Cash Flow share. And if we were to execute the $1.25 billion of share repurchases at today's price, a $140 per share, this range would move by $3. Now, obviously, if we pay more than today's $140, that range won't move as much, but I think those are the important metrics here. This is all done at constant shares at 51 million.
We do not incorporate them there. You can do the math, but again, threefold increase in free cash flow per share, 10% free cash flow yield, $3 of upside if we were to execute all the buybacks at today's stock price. Make one other mention on this slide. When you look at this, and when we talk about adjusted free cash flow, and when Terry talks about adjusted free cash flow, that's after tax, that's all in, okay? It's not free cash flow per share or free cash flow BG, before growth. We don't use that metric. This cash on the barrel head, cash that can be distributed, can be used for returns to shareholders. So that's different when you look at other free cash flow.
If you do all of that and you add and by the way, the midpoint here, many people are asking, the midpoint for 2026, 2027 uses the auction clear at $270. All we did was roll the $270 forward, the 2025, 2026 clear into 2026, 2027. $270 a MW-day. We did that. It's just a practice. We're gonna provide you a sensitivity later on EBITDA. Terry will walk you through that, okay? We also did it without the perfect and accelerate. We did it at the ramp schedule of AWS, all right? If you thought those things were gonna move forward or move higher, that would take you further up in the range, in addition to the share buybacks, okay?
Last thing, if you take the share buybacks, if you take some accelerations, put those in there, that free cash flow, that second number, that 10 that I was talking about, that 10% free cash flow yield, is greater than 11% free cash flow yield on a 26 metric, and as we're gonna show you later today, that's off of growing, stable, constant revenue streams, AWS, et cetera. We think that as we move to more of a contracted view or stable view of revenues and operating margin, that that lends itself to a lower cost of capital, therefore, a lower free cash flow yield necessary to run the business, so we're excited about the opportunity as we see it in front of us. Lastly, couple things. The share repurchase program, the $1.25 billion. Heard some comments from people.
It's authorized through 2026 , okay? What I would say to you on that is, well, that's generally how things are done when you get the approval of the board, and so therefore, that's obviously what we disclose. We also, I think, when was it? Earlier this year, March, April, I forget. Upped the share repurchase program from the original $300 million of October of last year, up to $1 billion, and we executed $900 million of that within six months. So don't read anything into that. It's just that you have to have the flexibility over time, and we can use that over time. We've done it faster this time. Who knows how fast we'll do it the next time? Next slide. Track record of unlocking value. So if you're gonna say something like that, you better back it up with a slide.
So that's why this slide is in here, okay? Look, when you come. If you've done restructurings like I've done in the past, I prefer to be doing what I'm doing now, quite frankly, after 10 years of bare-knuckle fighting. When you exit from bankruptcy, I think a lot of times you have the capital R, the restructuring, done. You restructured the balance sheet. But you have a little R, reorg, and optimization that haven't been done, okay? Now, when I talk about capital R and little R, our GC gets a little scared, and Terry gets a little scared because those have a different meaning when it comes to restatements in the SEC world. But that's not what I'm talking about here. I'm talking about optimizing the portfolio.
So when we came out, we've been focused on really two different types of unlocking value: transactions and managing the balance sheet/capital activities, okay? From a transaction perspective, we've done the Cumulus deal, we sold the assets in ERCOT for $785 million. We've done a number of other minor transactions. On the balance sheet, we've taken out our JV partners, we've simplified debt at the lower project levels, and we fortified our balance sheet, and that positions us for further growth and the further opportunities we have, okay? Why are we positioned in a strong position? It's because of our financials position in free cash flow and liquidity, our capital discipline with the cash that we're generating, okay? And we view our balance sheet as a strategic asset.
When I say that, Terry always reminds me, "Please be balanced about when you say that, because the first question Angie's gonna ask is, 'Why don't you lever up and use it for a cash buyback?'" So now you got to get a new question, Angie.
On your balance.
But the reality is, we don't need to lever up to do our $1.25 billion worth of buyback. We generate that in cash over time, okay? And we can use our balance sheet as a strategic asset. It doesn't necessarily need to be for buybacks. It can be, if we want to. It can be for growth opportunities, right? As of this date, we've basically done a lot of growth opportunities through the AWS contract, through the rest of it, either managing by cash through sales, cash generated, or basically being very capital light. So when we looked at the AWS transaction, we said, "We don't want to invest capital in data centers.
We want to sell power at better prices than we can get in the wholesale market." That didn't require any extra capital, that just required a lot of intellectual capital, okay? So we can use that balance sheet as a strategic asset in many, many different ways, but we're not going to commit to using it for buybacks today. You didn't even find that funny, Angie.
Warm me up.
Okay, good. That's good to hear. Next slide. So I used a version of this slide during the second quarter earnings, and I heard one feedback that it was-- I was using it to run for, you know, Secretary of Energy or something like that. But the reality is that there is a tremendous opportunity. Can you remind them back here? Thank you. The kitchen's right behind me, so it gets a little disturbing when they... There is a tremendous opportunity for the sector to step up where we are today to power the AI economy. And as I said, it's a great time to be in the IPP space. It's times that I haven't seen...
You know, we were talking, Greg and I were talking about cycles in this industry, and you see these boom-bust cycles, and they come and go, but this time it feels different. It feels structurally different, and that's why I say it's a great time. So we're seeing significant growth from electrification and demand. We're seeing queues that are almost entirely renewable. Now, I think we'll see a gas response, all right? But that's driving higher capacity prices, higher energy prices, and higher spark spreads. And that's led to a number of knee-jerk responses, in my opinion. One of those, and most of them are false narratives, if you ask me. One of those false narratives is that PJM is not working, and the result through this most recent BRA, Base Residual Auction, signals that, the $270 a MW clear.
I would argue or contend it's just the opposite. PJM is a market that works, and it is sending a price signal, and the next auction that clears in the fall will continue to do so. And probably the one in June of next year, I think it's June, of next year, will do the same, given the supply-demand response. That is a working, functioning market. And by the way, over the past two decades, PJM has brought to bear more than 60 GW of new gen, okay? Now, it's been offset by retirements. So one of the other false narratives is that you hear from NGOs, consumer advocates, policymakers, is that same argument that it's not working and it's driving up higher prices. By the way, they're the same organizations that their policy decisions have pushed all that off the grid.
So they push everything off the grid, want to go to renewables, and then say, "Well, the prices should remain low." That's not how it works. They've opposed pipelines, new build generation, transmission, et cetera. So we're at an inflection point where gas is going to have to solve the problem, gas generation and the new queues, but I think it will, okay? Some of those policy decisions have also been exacerbated by price capping and the price mechanisms by which the IMM is implemented, all right? All of which is leading to this confluence of why current generation is of high value. And that's leading us, by the way, to look at how do we reinvest in our current plants. It wasn't but 15 months ago, 16 months ago, that when we emerged from bankruptcy, people said to me, "The gas fleet is worth zero."... Okay?
And when I looked at the economics, when we looked at the economics on a pro forma basis, it's close to right on some of the assets. They were marginal on some of the assets. We always thought they were of value. But now we're looking at reinvesting in those assets. Not a lot of capital, but you got to put capital back in because we've gone from running our fleet and doubling the capacity factors, particularly in the second quarter of this year on the gas fleet. All right, so knee-jerk reactions, false narratives. How should we respond as a sector? I think we should look at the economic development opportunity by which this all brings. You heard me talk about data centers bringing huge economic development to regions. It is a win-win for everybody. I just think we're losing sight of that.
It's a win-win for generators, T&D companies, load-serving entities, local communities, basically all the stakeholders, because it's billions of dollars of investment. That's. Every time you turn on or listen to CNBC or read something on the Journal, it's about the big four and the, you know, the hyperscalers going to spend a quarter trillion dollars a year to go after this, well, that all needs power, and that's what our AWS deal does, so we're well positioned because we feel as though we're at the intersection. We've done the first deal. We get to capture upside in capacity and spark spreads, and then we have protection through some of our long-term contracts, the PTC, et cetera.
Lastly, before I turn it over to Terry, for those of you that are new to the story or semi-new to the story, I'm not going to bore you with going through all of our assets. You can go in the back in the appendix, and there's a cut sheet for every one of our assets. It, you know, walks you through what they are, as well as some additional modeling detail. But our portfolio, which is principally in PJM, as I mentioned earlier, just by the numbers, we're 10.7 GW of generation across 12 facilities and 1,900 employees. 50% of that generation is carbon-free, coming off of our Susquehanna nuclear plant.
One of the numbers that's not here, but I'll leave you with before turning this over to Terry, is this: when we exited, emerged from bankruptcy May seventeenth, 2023, our enterprise value was roughly $4.5 billion. Our enterprise value today is roughly double that, $9 billion. So where do we go from there, right? It's a question. You've doubled the enterprise value. Well, I think if you go back to the original discussion of free cash flow per share, we're tripling free cash flow per share, and that does not include all the upside of the AWS contract on the come. It has midpoints using the 2025, 2026 curve, does not incorporate our share repurchase program, et cetera.
I think that there's a tremendous opportunity in the value of Talen, and that's why we continue to think the highest and best use of our cash flow generation, Terry's going to get into the capital allocation, is to buy back shares. And as I mentioned earlier, we have the strategic asset of our balance sheet that we've grown into and will further grow into as we add further contracted revenue streams and implement the AWS contract, which should be highly leverageable, given that it is a double A credit on the opposite side of that revenue stream. And I think we have the opportunity to add more to it by leveraging what we did in that first contract against more. Many of you have heard me say, I hope somebody else in the industry announces a deal.
I just hope we announce our second before anybody else announces their first, so with that, I'm going to turn it over to Terry.
Thanks, Mac, and good afternoon, everyone. Turn to slide 10 on the deck. I want to do a couple of things, today. First, provide adjusted EBITDA and adjusted free cash flow guidance for 2025, and then secondly, provide a preliminary outlook on our 2026 metrics. As you can see here on the chart, our 2025 guidance has a adjusted EBITDA range of $925 million-$1.175 billion, or a midpoint of $1.05 billion, and our adjusted free cash flow is $395 million-$595 million, or a midpoint of $495 million.
Our 2026 outlook has an even higher adjusted EBITDA range of $1.13 billion-$1.53 billion, or a midpoint of $1.33 billion, and adjusted free cash flow of $535 million-$895 million, or a midpoint of $715 million. As you can see by the arrows on both sides of the chart here, our adjusted EBITDA, our earnings potential grows at a significant compounding annual growth rate, well in excess of 30% over the next few years. What's even more interesting, and to add on to Mac's comment earlier, is our adjusted free cash flow almost doubles that growth rate. We think that's really led by a few things.
Number one, it's just our ability to convert a significant amount of our incremental earnings into cash due to very limited growth CapEx. As Mac alluded to earlier, a number of other peer groups will talk about free cash flow before growth, but then they'll come back in, and they'll say, "Well, I may have $700 million or $1 billion of growth CapEx, which you guys should just ignore." We don't do that. That's not our approach.... We believe that these growth rates, along with other potential catalysts, provide a compelling investment opportunity with a strong growth profile. So the size of our ranges are dependent on several factors. First and foremost, is sensitivity to power prices and spark spreads. However, the good thing is our commercial hedging program does provide risk protection around that.
We're nearly fully hedged for 2024, about 2/3 hedged for 2025, and just over 30% hedged for 2026. Chris will talk more about our commercial hedging program in a bit. Second, and Matt noted this earlier as well, the 2026, 2027 PJM capacity auction is also a variable when you're looking at our 2026 outlook number. We've made a very simplifying assumption and just taken the last capacity auction clear of approximately $270 a MW day, and extended that forward for the 2026, 2027 outlook. There's a potential for this to be higher. You've seen a number of people write about it, discuss it. We're not gonna go into too much detail on that. Chris will talk about it a little bit shortly.
But the other thing we've done here is we've given you a sensitivity or a rule of thumb that you can sort of toggle those results by. So a $50 /MW day change in the 2026-2027 planning year results impacts our fiscal year 2026 adjusted EBITDA by approximately $50 million. And so you can lever it up and down with that simple rule of thumb. So turning to slide 11. The strong growth profile that I previously discussed around our earnings and cash flow provides the opportunity to continue to execute on our capital allocation plan. Since announcing our repurchasing plan back in October of last year, I'm pleased to say that we've returned over $930 million of capital back to shareholders. This has reduced our share count by approximately 14%.
Continue our commitment to shareholders. Earlier today, we announced the reloading of the share repurchase program to $1.25 billion through the end of 2026. We continue to believe that buying back our stock is the best, highest, and best use of our capital. Our current share repurchase program is supported by unrestricted cash for approximately $700 million on the balance sheet at the end of August, as well as an excess of $1.2 billion of adjusted free cash flow generated over the next two years. As part of our go-forward capital allocation strategy, we've also provided a targeted allocation on this slide. You can see here we plan to distribute approximately 70% of our adjusted free cash flow to shareholders as we move forward. Moving to slide 12.
We've got significant margin growth over the next several years, with increasing stability across the board. Our projected gross margin increases from 2024 to 2026, and more of it will come by stable sources backed by large creditworthy entities. That percentage is expected to continue to increase past 2026, as we continue to ramp up the power that we serve at our AWS data center. These stable, stable sources of earnings and cash primarily come from the revenues associated with the AWS contract, capacity payments from PJM, and earnings backed by the nuclear PTC. This should reduce our liquidity needs, increase our hedging flexibility, which both in turn lead or lend themselves to a lower required free cash flow yield for the enterprise.
As we think about demand for our equity, following our uplisting to NASDAQ in July, Talen is now more accessible to a broader number of investors and a broader investor base. We have or hope to become eligible for various stock indices as we move forward. We expect Talen to be incorporated into various indices run by S&P, CRSP, MSCI, and the Russell, as soon as the second half of September. And we also anticipate this will drive significant incremental demand from passive index funds. Talen has a versatile value proposition and could qualify for additional value, growth, and other industry-specific indices as we move forward. With that, I'll turn the presentation over to Chris, and he'll talk about several market tailwinds and our hedging strategy.
Mm-hmm. Thanks, Terry. Thank you all for coming. I'll start macro, get a little more granular on PJM, and then double-click into Talen, more specifically. In my nearly 20 years in the power space, I have not seen the structural demand growth like we're witnessing today. That's across the U.S., but more acute within the PJM footprint. Primarily driven by the three key factors you'll see here at the bottom of the slide. Data centers, we all know and love that story. The return of domestic industry and manufacturing. We're seeing, you know, on the heels of the data center play, the chip plants, the processing plants, there's a reshoring initiative. So business is back on that front. And then lastly, this general trend towards more electrification, EVs, chips, home heating.
PSEG put out a fully digitized kind of port in New Jersey that they've that's out there. So more of that across all sectors, driving large electrification. Looking at the chart to the right, specifically at PJM and the related demand growth expected to materialize there, largely driven by the three factors we just mentioned: data, business, and electrification. PJM's own load growth model has increased nearly 20 GW from just a couple years ago. So massive growth relative to history, a dramatic increase relative to their own studies of just a few years ago. Flipping the slide, looking at some market dynamics. The supply and demand fundamentals are increasingly getting tight. We've previously discussed the demand side of the equation and the robust growth we're experiencing on that front, but similarly, supply is, it has its own set of headwinds to deal with.
Increased retirements, lack of new build, kind of the primary drivers driving that tightness in the supply. PJM reserve margins are falling, and looking to the left here, declining reserves create a trend line that is directly going one way, downward. This makes existing resources increasingly valuable, with an accelerated focus on grid reliability. Through RMRs and historically clear auction prices, PJM is sending a very clear signal: they need generation. The latest BRA clear is indicative of just how tight things are getting. Add to it, continued expansion of spark spreads in the east, and Talen is well positioned to benefit from these externalities. We touched on it, we weren't going to get out here without getting more specific, but the 2026-2027 capacity parameters have been released. Of note, the 2026-2027 load forecast and internal reserve margins are both increasing relative to 2025-2026.
Unknown, however, are the supply changes, right? I think new supply can come in the form of a few things. New build, right? There's the one gas plant that's kind of been out there, Trumbull, 900 MW out west. Potential for announced retirements to potentially come back. It's a small list, but that does exist. Demand response is a bit of a wild card, right? We can potentially see more demand response get offered into the grid or change on their bidding behavior itself. And then lastly, uprates. So we think in the aggregate of all those things, potentially only bring back two to three GW of additional supply. So basically, while the sphere of unknowns has shrunk, given the steepness of that demand curve, these small changes in supply could have a dramatic sort of range on the potential outcomes for the auction.
Lastly, looking at our commercial hedging philosophy, you'll see on the left here a diagram depicting the numerous variables contemplated in determining the most appropriate, and more importantly, the most effective hedge. We do have established hedge targets, and they're just that, targets, which do provide direction on the amount of volume we aim to de-risk during different periods of time. About 12 months, 60%-80%, the following year, 40%-60%. Beyond that, we're constantly monitoring, evaluating, and balancing some combination of our views on markets, market liquidity, geopolitical factors, regulatory risk, credit risk, just to name a few. Our commercial hedging strategy is pragmatic, not programmatic. In saying that, we can be opportunistic in our buying and our selling. Given what we own, our natural position is long the spark, or dark, or crack, or cork.
We own the power and we're short the fuel, right? So this gives us the freedom to be a little less formulaic when initiating some of those hedges. In times of extreme market volatility, it may make sense to do more or less of said hedging activity. Looking now at the bottom right of the slide, you can see our commercial hedging program at work in a table with sensitivities assuming various power price changes. The implications of the nuclear PTC, some other strategies, proprietary strategies that we implement, help us achieve some asymmetric results with more favorable upside skew, while also providing adequate downside protection.
For example, you'll see in 2025, a $10/MW increase could lead to nearly $200 million of gross margin increase, while a $10 price decrease results in only a potential $50 million margin decrease. In summary, our hedging strategy preserves margin, provides cash flow stability, and generates upside in a variety of market conditions, as evidenced by our outsized commercial contributions since starting this journey almost a year and a half ago. I'll now turn it over to Cole to talk all things data.
Thanks, Chris, and good afternoon, everyone. All right, so I thought we'd start this section kind of doing a bit of a recap of the deal we did back in March with Amazon, especially for folks who are maybe newer to the story. So the deal with or the transaction with AWS was really, we sold our campus, our data center campus, for $650 million. But importantly, it's not just the sale of the asset, but it's also the long-term PPA that we contracted with Amazon along the way for power from our Susquehanna nuclear facility for the life of the plant. And that goes through 2042 currently, and there's options to extend that as the nuclear license is extended beyond that by the NRC out to 2062.
The Susquehanna PPA has up to 960 MW of direct connect, carbon-free power available to the data campus, and Amazon has the ability to access this 960 MW immediately, or at least as fast as they can develop the campus, and then there's a backstop for Susquehanna on minimum contractual ramp obligations for Amazon to upsize the PPA by, or their payments by, at least 120 MW, starting in 2025, and increasing at 125 or 120 MW increments from there. Each of these increments or these 120 MW tranches, as we call them, are fixed price over a 10-year period, and after that rolls off, they're repriced at a market-based mechanism that has an embedded premium still in it.
There's also another contract for additional revenue streams from the carbon-free energy sales for separate power that Amazon's also paying under this agreement here, and then turning to the campus itself, we did this transaction in March. That was, what? Six months ago now. Amazon's publicly stated through various processes up on site or up in the township that their stated goal is to develop the campus fully within five years. Right? So a lot faster than the 120-MW contractual ramps that they've committed to, and I think the evidence so far has been very overwhelmingly positive to us on site, so there's been a couple of milestones met already. I'll just go through the ones up here on the screen, so in May, the township approved some zoning changes, which enabled no...
Which enabled data center use by right, which effectively means no more conditional use permits and discretionary permits for every building, going forward. In June, they got, Amazon received the first building permit to fit out the existing data center shell that you can see in the picture there. It's the shell that we built and then ultimately sold to Amazon. Then in July, Amazon received approval for the master site plan, which was, you know, 14- 16 buildings, the entire infrastructure for the site, that unlocks the full campus development. And then just last month, Amazon released the $300 million of the $650 million of the initial purchase price, which was held in escrow, pending milestone development or development milestones that were achieved.
And so, you know, we're looking forward to continue to partner and support Amazon here as they continue to develop the campus at pace. So let's turn to the financial impacts. And, for folks who are checking and doing a red line, this is the exact same slide that you saw back in March. So, we thought it was best to show. It's the same deal in March as it was, as it is today, right? And so we didn't wanna confuse folks by looking at different ways to compare. We'll talk through some sensitivity on that in a second, but it is the same slide.
And really, just to reiterate for folks who maybe weren't around the story in March, what you're seeing here is the EBITDA growth or the EBITDA impact relative to the base nuclear plant selling into the markets. And so really, it's the Amazon deal, less some kind of reference price. And the reference price we used in March, and we're showing here still, is the PTC floor. And so for every dollar that you want to increase that reference price per MW h, for every 120 MW tranche, you're gonna see a $1 million decrease in the net impact. And so the bar, again, that the top line stays the same as you fluctuate market price or capacity price.
Obviously, those have changed a bit, and that goes up and down, just the incremental delta may shift here. And we think it's helpful to look at the top end of this range or top end of the bar here, because we're looking at 120 MW tranches built out to the 960 over a period of time. That shows you that that gets you to the top end of the range here. As Mac said earlier, we're looking to perfect and accelerate. And so on that acceleration, if we meet that Amazon timeline of five years, you can meaningfully move that right-hand bar to the end of this decade versus the middle of the 2030s , right? And so that creates, you know, huge value upon accelerating there.
So hopefully that's helpful for you, all to see, you know, that we think it's a really exciting deal. I think, many of you also agree with us on that, and it's why we're going to, continue to work hard to leverage this in other assets, and I'll come back to that in a second. But first, as Mac said, we'd like to talk a little bit about the ISA and FERC process ongoing. I'm sure most of you are tracking that at this point. Again, just maybe level set with everybody in the room. What an ISA is, it's just it stands for Interconnection Service Agreement, and it's a three-way agreement between parties that every generator, by the way, has to have, right?
It's the generator having a deal with the system operator and the transmission owner, all three of them. And so in this case, it's Susquehanna, PJM, and PPL. And we need to amend this ISA to support the generator, in this case, Susquehanna, providing power directly to the data campus without using the grid. That's important. And for folks kinda newer to the story, when we talk about serving the data campus in the context of the ISA, they call it co-located load, right? So you'll see the term co-located load a lot. I think it's helpful to first start with currently what's in place. We already have amended this ISA to allow for 300 MW to support co-located load. We did this two years ago.
PJM and PPL both reviewed the reliability aspects, signed off on it, and yes, FERC approved it, right? What we did as we were going through this process over the last, you know, two years, upsizing the campus to 960 MW, we submitted this into PJM for review. And what they do is a necessary study, which is effectively an analysis of the entire grid, the network in the area, to make sure there's no stability concerns or reliability concerns. And they, along with PPL, who reviewed the results, concluded that we can go up to 480 MW directly to the co-located load immediately with no stability or reliability issues.
And if we address one minor stability issue that was noted elsewhere in the region, we can then go up to 960 MW. And that concern was, you know, basically alleviated with a less than $3 million capital upgrade that we've already paid to do, and that work's ongoing. And when that's complete, we will be able to upsize that 480 to the 960. So PJM submitted that to FERC, that amendment of the ISA to FERC early in June, and they submitted it with one additional inclusion that both PPL and PJM thought was helpful to include, and we agreed, which are additional enhancements to the reliability, to safeguard the system, to make sure we weren't inadvertently touching the grid, pulling power from the grid, et cetera.
And that's what the Schedule F that, that folks maybe have read about really outlines. You know, once we submitted the ISA, as folks know, there was a few folks that protested, a few, you know, fairly large folks that protested our ISA. They're folks that are not related to this agreement or really anything in our jurisdiction here. FERC. And that obviously created a lot of, I won't say noise, but a lot of discussion, a lot of dialogue that I would say has probably reached the national level, right? Our ISA was reviewed by FERC, and ultimately, in early August, they issued what's called a deficiency letter, which might sound like a negative thing, but all that really was, was a request for more information.
And what they said is, "We want more information re, regarding these reliability safeguards that you put in place, and please explain that a little bit more." And so what PJM did, working with PPL and us, they this past Tuesday, filed a response that gave a pretty fulsome outline of why those safeguards were helpful, why it protects the grid even more, and why this is actually reliability enhancing versus maybe some the other way, which is some of the plans from the protesters here. So that leaves FERC with now a 60-day window to review and ultimately opine. That 60-day window expires on a weekend, so go to the next Monday. We're looking forward towards March, or excuse me, November 4th, for a response from FERC on this issue. And we're optimistic.
We continue to be optimistic that once FERC reads PJM's response, goes through it fully, and understands the setup we have, the unique characteristics here, that they will move forward and approve the ISA. Separately, FERC also issued what's called a technical conference or set up a technical conference to discuss the broader issues that have clearly been bubbling up since we've announced this, since the protest happened. And the industry is really engaging about: how does this look for different situations? And each transmission owner, each generator, everyone's gonna have unique circumstances. And so how do you put a broad framework around that? And so that's really what we expect FERC to be talking about at this technical conference. Notably, it's not involving our ISA.
They actually bifurcated it from our ISA, and they actually can't talk about any specific proceedings in this conference. We will get more on the agenda here probably in the next couple of weeks as to exactly what FERC wants to talk about and how they talk about it. But we are going to participate in that, as we expect a lot of the main players and parties of interest to do so. And Mac said this on the earnings call, but we really think that a rapid resolution here is key. It'll if they don't, and this kind of waffling continues out there, it's gonna chill investment, it's gonna chill growth.
And ultimately, our deal might not be the deal for everyone or every generator out there, but we do think it is one solution of many that can ultimately solve the problem. The data load's coming, it's already here, it's already moving at scale, it's already expanding. And we think our deal and others like this will continue to maintain reliability for the grid and will continue to have no adverse impact to ratepayers. And so ultimately, you know, we're very, we're looking forward to the outcome of this technical conference as well. All right, so I'm gonna end here with just a quick discussion on the Talen platform and what we think is the first mover advantage, as Mac alluded to in the entry there.
And really, this was developed over four years of time, and, you know, kind of go through the different elements there. And just to cut to the chase, we're not announcing any data center deal here today. That's maybe for another day, but I'm sure the question is gonna come up. But really, I think we've built a valuable platform, really around two different themes here. One is, you know, we understand what's important to hyperscalers, and that might sound, you know, pretty obvious when I go through the pieces here.
But as we've contracted with AWS, and as we've discussed and had dialogue with numerous other parties through the process here, it's really about the speed to power and the speed to market. That's first and foremost, but it's not just about the time to getting to that, it's also, too, about the scale. Can you get to a GW? Can you get to something more? How fast, how big is really the question. And hyperscalers aren't wasting their time on... It used to be 50 and 100, and then 250, 500 MW. Now we're really talking giga scale for these campuses and where the focus is starting to shift for these AI compute loads. Third thing is just 24/7 reliable power. I think that's pretty intuitive as well.
Low carbon or no carbon in the case of Susquehanna. I think an under appreciated aspect is having long-term PPAs for the hyperscaler to have both attractive prices, but also price certainty. I think that is very key as they think about investments. There's $250 billion of investments these hyperscalers are making per year into AI infrastructure and compute. They need to have line of sight that the power's there and what that price is. The other thing as I mentioned earlier is we've built the IP as an organization. So we've done all the technical development. We've figured out how to connect the infrastructure directly into a nuclear facility and all the complexities around that.
We've dealt with all the zoning, we've dealt with all of the necessary studies, all this, the technical analysis, to make sure we're on sure footing, and we know how to do that and replicate that faster and faster now that we know kind of the which way to pivot, depending on the circumstances of each asset. We also have engaged with a number of stakeholders, PJM, PPL, NERC, FERC, state officials, local township supervisors. We understand how to navigate that system pretty well at this point. As Mac mentioned, commercial contracting, right? So we obviously got to a deal with Amazon that took quarters, not weeks and months. It took us quarters to get there.
But we also were dancing with a number of other parties, and we understand what similarities each of them had and what differences each of them have. And so how are different constructs, potentially, how would they work for different assets? I think we have an advantage on that. So bottom line is, based on our learnings through this process, at least at Susquehanna, we're we believe that we're able to now leverage that across our asset base. Maybe it's a little bit different, and it's unique for each asset, but we do think there are future opportunities, and we're certainly working through that. And you know, I think more to follow here, in the coming period of time. So with that, I'm gonna stop and turn it back to Mac.
Thanks, Cole. I got this vision of you dancing with counterparties. It's a little awkward. He might wanna not use that one next time. In any event, Cole's done a great, great job. One of the things that I think gets underestimated with this, there is a lot of technical capability, but there's a commercial aspect of this. Cole and I have been on calls just this week with parties that are interested in trying to develop data centers, you know, with power. And it's become very evident to us that, you know, we're at an inflection point in power, trying to solve this need of the growing demand and this AI economy.
But I also think for those that keep track and, and look at the data center space, and look at the NVIDIA space, and, and the derivatives of NVIDIA, that it's at an inflection point too, because the old data center model is evolving rapidly, very rapidly. It used to be plop down in Northern Virginia, sell 40 MW, plop in a-- recycle the capital, and plop down another box. You heard Cole talk about it's moving to these gig centers or multi-gig centers, and that's where it's going. But there's also a difference, which is data center operators and hyperscalers are also getting smarter because it's that old location, location model, is power, power, power model now, and they've got to get real smart on power.
And so we learned a couple things going through the commercial activities, which were, you know, data center operators want five nines of reliability, 99.999% reliability. And when you ask why, there's no answer. Because. Well, there is an answer. It's because that's what we've always got. Okay? But what's your business interruption cost? The old model was, I plop down a 40-MW shell, and I take 20 MW. The utility's on the hook for keeping the other 20 MW. I don't pay for that, okay? Maybe the demand charge, but not the energy charge. Well, that's very inefficient, as you can imagine, for those that are in the power sector, like having to reserve 40 MW at all periods of time, but only using 20 MW and getting paid for it on the energy, you're burdening.
So you're going to see a significant change in the way that data centers have to move to integrate power, and I think that's one of the things that we understand and can bring to the table when we talk about either co-located deals, front of the meter, behind the meter, however the next evolution is, okay? We understand that, and I think that's. We describe that as IP. I don't know that it's IP. I don't know that it's IP five years from now as the market evolves. But having that understanding, I think, gives us a first mover advantage to be able to leverage that. We like to say Talen's powering the future, and we're uniquely positioned to do so. You heard me talk about it.
Strong cash flow allows us to do the $1.25 billion, the fleet, first mover AWS growth, that disciplined track record. We continue to say that that's table stakes. And as I mentioned, we've got a balance sheet that we're going into. It's a strategic asset. I think that sets us off on a good platform on the heels of everything that's been done. And so with that, I'm going to stop and get some more energy in the room with questions.
Okay. Thank you, Mac and team. Now it's everyone's favorite time, time for Q&A. Also, time for snacks. So there are snacks and drinks on either side if you'd like to stop and grab some, or you can wait until the end of Q&A. But most of you have received a laminated number for those of you in the room. The numbers don't mean anything aside from enabling me to call on you without needing to awkwardly try to describe you. So for those of you who are virtual, it'll be a bit easier. Just bop your questions into the Q&A function in Zoom, and I will get to them as well. All right, let's see. Kicking things off. Let's kick things off with number 19 over there.
Does this work?
Yep.
Hey, Nick Campanella, Barclays. Thanks for doing this late in the day.
Thanks, Nick.
Thanks. So, you know, how does the recent capacity auction being elevated, and then, you know, what we're seeing in terms of December potentially clearing higher, how does that change your discussions on price, and how you, you know, you're thinking about new data center signings, and how does it also change, you know, customer motivation?
Well, I'll start. I think I'm gonna go in reverse. From a customer motivation, I don't think anything's gonna stop the insatiable appetite of the AI data center demand. I don't know that it's exactly the forecast that you see out there, because there's going to be changes. It looks like queues, right? People used to get in the LNG queue at FERC, people getting in the queues at PJM to build new stuff, and so some of that will go away, because this is just on the demand side. There will be cooling improvements, which will reduce demand. There'll be improvements that happen in chips, and you know, right now, the improvement that's happening in chips is more power, but more cooling per square foot. So you know, there's a lot of dynamics that go into that.
So I don't think the demand side of things changes, I just don't know that all the forecasts will become realized. But it doesn't matter, because if you have 50% of what's forecast, you still have reserve margins in the midsection in the out part of this decade that Chris showed. With respect to pricing, you know, it's really interesting. It's a little bit of a self-fulfilling... You know, I don't like using words like, you know, we had this milestone deal with AWS. It's more like an inflection point where everything started to rationalize its descent. But recall that all the forward curves and everything started to respond after doing the deal.
And then it was unique because we were at a you know sell-side thing, and someone said: "Well, you know, what does this look like mark-to-market now that all the prices went up?" I said: "I don't know. You know, it's not as much in the money." But we got 1,300 more MW in new, and another 9,000, you know, 8,000 MW of other stuff. So, but what will it do going forward? Obviously, when we did the first contract, it was priced off of a view of market capacity and energy at the time, with certain parameters that are somewhat confidential and private. But there was a basis. The basis is just now higher, and I think you'll see deals get done at higher.
And then eventually, once this demand starts to get solved, now I'm talking a decade from now, the prices will come down, probably. But that's not going to happen for a while. I hope that answers your question, Nick.
Yes, thank you very much. And then just you're highlighting a lot of demand for the stock with the index additions. And, you know, obviously, we have this, you know, December 2024 catalyst with the auction clears pretty high. With the buyback being through 2026, you know, why wouldn't that be, you know, more accelerated to be as soon as possible?
We don't want to let you know when it's going to happen. I didn't mean to be smart about it. Go ahead.
No, it's fine.
It's the truthful answer, but there's truth in the sarcasm. But yeah.
All options are on the table for the buyback?
Correct.
Thank you.
Maybe the other thing, Nick, maybe going back to the first question. You have to keep in mind that these data centers are long-term assets, right? They're real estate investments. So even though you may have one or two high-capacity clears, right, this is a, you know, 20+ year investment for these folks. And so they've got to look on the horizon, and think about what that cost means for the life of the asset.
Can I make one other comment, though, just so everybody's aware? We had a $1 billion authorization back in... When was it again? I, when you, you're younger. April. But there was periods in time by which we couldn't necessarily at that time, point in time, we were in the process of signing documents with CPS to sell the $785 million ERCOT assets, and so we had MNPI. And then you get into the quarter, and you get the blackout period where you can't do things. So there's some constraints on things. We would have executed because people say: "Well, why didn't you execute that $900 million earlier?" Well, the tender and then the bilateral. We would have if we could have. So maybe just another data point.
Thank you. Yeah.
All right, let's go with number 110, Angie, since you got some mentions from Mac.
Exactly. So first, I mean, just like taking a step back, you seem to have a lot of very conservative assumptions across basically all of the metrics, which strikes me as strange for a company that, as you mentioned, repeatedly comes out of restructuring. That usually means you will try to talk up your story. So maybe it's just your style, but again, you seem to be overly conservative versus your peers. So that's number one.
It's the CFO's fault.
That's Terry's. Okay.
I'm kidding. You want us to answer the question? Look, there's a, I think it's a motto at Franklin and Marshall, but Ben Franklin once said: "Well done is better than well said." We'd rather do stuff and then tell you about it, and, you know, you've heard us do that. I'd rather put out something that is based and grounded, so that you can then make your own assumptions from there. That's why we provide all the sensitivities, and you can look at it and do those types of things. I appreciate that you think that they're conservative and that you view us that way, but we like to set bars and meet them or beat them, and not, there's no need to talk, you know, talk our book. And he's conservative. A&M.
So, um-
People don't know what that means. He's an Aggie.
Okay, so then, maybe, a bit more about this potential accelerated, build-out of the Cumulus site. So, we obviously saw the statements from AWS, and just wondering if there's any constraints or if we want to think about the, you know, how fast this site can be developed. Is there like a technical limitation? For example, how many MW per year can be absorbed?
I wouldn't say there's a technical development. And first of all, I should put the caveat, you know, we're not going to speak for AWS, and ultimately, they drive the development schedule. But look, we built that shell that you can see in the picture, or I think that was on the previous slide, and sold that to Amazon. And as I said earlier, they're obviously looking to fit that out, but also accelerate and build out the campus. And so that means they have to buy equipment. So I do think there is a physical limitation in the near term, let's call that in a year or two, for all the equipment to, you know, long-lead equipment to be bought, additional transformers, additional substations, et cetera.
So I think in the nearer term, year or two, there's probably a structural limit. But beyond that, it's really up to Amazon. How fast do they want to buy? Three substations worth of transformers or six, four buildings, or eight, or 10, or 12 at once, and have that come in the next, you know, year or two. And then ultimately, do they have the labor to build all that concurrently, right?
Then just one last one. You didn't mention anything about the coin, the coin business. So I was just wondering if you could comment about potential conversion of that business into a data center.
Yeah, I think, you know, we're gonna stay kind of on the line that we were at at the Q2 earnings, which is, you know, we're working on different options there. I appreciate we've probably been saying that for a bit of time, but I think more to come on that, but not today.
Yeah. I think, Angie, just the simple answer is we're, we remain committed that it's not part of - and as we've said, it's not a strategic asset for us. It's just, you know, until we decide what to do with it, there's not a lot to say.
... All right, I'll do 111 , and then 113 .
Thanks.
Oh, press the button, though.
Yeah. I know you touched on it a little bit, but can you just maybe put a little more color around the other 30% of the free cash flow allocation? I think on the last call, you mentioned maybe some things to do with the RMR units, but is there anything else there in terms of, growth CapEx or, yeah, just non-buyback capital allocation?
Maybe a couple of things on that. Michael, appreciate the question. So the 70% is a target. It's not a hard and fast rule. It can be above that, it can be below that, right? You know, when it comes to the excess cash that we generate and it comes to, like, inorganic opportunities, I think the one thing that I'd be cautious on is, you know, this team up here has been through cycles. We know what cycles look like. You know, if there was a value proposition out there that we thought had, deep, deep value in it, and was a strategic fit for us, we would go out and take a hard look.
But the good thing is the balance sheet and liquidity that we have, and the ability to have excess capacity gives us options to do that. It's not like we're going out there and, you know, looking to buy anything and everything, and we would never do that. That's ultimately not in the cards for us. We get dinged. I mean, I'll be honest, like, I talked to the rating agencies last week, and they had nothing but good things to say about the results, but then bad things to say about, "Well, you guys are too small, you're not diversified.
You know, why don't you go out and buy this and buy that?" I said, you know, I told all three agencies, "We're not going to go out and do a transaction just because, you know, there's a transaction to be had. It's got to mean a lot of value for us at the end of the day." So that's how I would sort of characterize the remaining 30%, and it gives us dry powder. It gives us, you know, the opportunity to do things, whether maybe it's on the equity side, maybe it's elsewhere.
The other thing is we have $690 million-$700 million of cash on the balance sheet. It's in the appendix. We're generating over $1 billion. We have this balance sheet that allows us to do a number of things. So... But people always ask us, "What's the capital allocation policy? Give us a number." It's like giving a hedge number. And that's why Terry was very careful to say it's a target, because we can flex that up or down. If you look at the position that we're in, we're in no better position than we've been in a long time. We have revolver capacity, we have first lien capacity to do trading and hedging, okay?
We manage that, look at that daily, weekly, get together as a team, but we manage it on a daily basis. And we don't need to hedge as much because we got the production tax credit to the downside, and we're having this growing, stable revenue stream from AWS. And so when we look at all of those things, we have a lot of flexibility to do right now. But as you look out into the market, you know, the M&A market, by the way, when I'm talking M&A, I'm talking assets or what we might do for, you know, small amounts of growth, it's kind of locked up right now. Everybody thinks their stuff's worth a lot of money, and everybody else says it was one auction clear. And, not, you know. So we're out there.
We always constantly evaluate, and we look at whether or not we're going to sell assets or maybe add something that provides a growth opportunity. But we like where we are because we have something that other people don't have. We have the ability to perfect and accelerate the AWS contract, pull that forward, and that is growth. That is real cash flow growth over time, that is seeable, and all we got to do is work it. Other people are looking at it, and basically, when you buy an asset, you're specing the market, right? You're taking a view on the market, and the market's going to be better from here. We don't have to do that. All we've got to do is convert MW from going to market to going to AWS, and then figure out how to do that again at more locations.
That's something that others don't offer, and that's with a double A credit on the other side. There's a whole bunch of other things, and so that that gives Chris more flexibility in doing to what he's doing, but it also gives us more flexibility to go look at what other alternatives could we figure out to add on or cut off. We did so with the ERCOT assets, right? We're looking at what we do with coin.
Great. And then my second one was just around maybe risks around the PJM auction. So I know you use the $270, that's kind of a roll forward, but you gave some, you know, drivers. So what gives you confidence that, you know, you're using that as a marker now, and you guys have been around a long time. We cleared at an all-time high. These things can be unpredictable. And then on the flip side, the risk if it prices materially higher, the political blowback around that.
Can you talk about bilateral?
Sure, yeah. So, you know, a lot of debate on, given the historically high clearing, is it prudent to carry that forward? As you know, expectations of clearing cap is viewed conservatively. You know, anecdotally, I'll tell you, there's some bilateral trades that have gone through. The capacity bilateral market's pretty opaque, pretty thin, to be frank. So, you know, those are being priced north of the clearing, right? The, you know, 300s, 330s are a couple of bilateral trades that I've seen. Again, just giving us confidence that the 270 was, you know, a safe number, given the last clearing. And again, the sensitivities provided can allow for some toggling above or below that. But, you're right, there's a ton of risk to the downside, ton of risk to the upside. It's PJM capacity.
It's a bit of a beast in and of itself, so yeah, the 270 assumption was well defended, I'd say.
Yeah, and maybe just to add Chris's comments too, this upcoming auction is going to be, given the parameters, the steepness and the curve is a very unique situation, this upcoming auction. I mean, I don't know if we've ever seen a curve that looks like this, right? Higher reserve margin requirement, very limited, actually, a reduction in some of the import capacity to several zones, right? That makes for a very, very steep curve. And so ultimately, and Chris alluded to this earlier, when you think about the main variable, it's really going to be on the supply side. What happens with supply? You know, does DR come back in the market? PJM doesn't have a massive demand response market like maybe ERCOT does, and so there's some limitation there.
But supply is going to be the variable that I think we'll have to see when the December auction comes, and quite frankly, even the auction in June of next year.
But I, I think, Michael, to answer maybe your political question is maybe. I don't want to go down history lane here, but, you know, how we ended up here is because of battling over what these rules should look like, and that's not PJM's fault. PJM's been trying to get these auctions and get them on a three-year forward look, which is what they're supposed to be, right? So that you have a signal. I think PJM's put forth, you know, don't always agree with PJM, but they put forth, you know, at FERC, some of which were opposed by the IMM, and ultimately, the FERC agreed with the IMM and shut down some of those rules, okay?
They've been trying to run these auctions, and they're doing. You know, like I said, they've brought on 60 GW over the last couple decades and made a functioning market. That's a good thing. But the reason, you know. And I do think politically, you'll see the rhetoric heat up when the auction clears high. And that's because you're going to hear more people like the letter that came out of the consumer advocates say people are paying higher prices. Well, that's because policies matter. Like, if you have policies that push generation off the grid, I don't know what you think the response is going to be. I mean, it's simple to be. You know, it's like, is that micro or macroeconomics? I was an engineer.
That's macro.
Okay. But, you know, it's like, I don't know what people think the response is going to be. You push stuff off the grid, the demand keeps growing, and people have been able to, you know, survive that, or the grid's been able to survive that by and keep prices low because demand has been going away. And I think what's different now is that demand is coming, and if PJM doesn't get it right, if PJM doesn't get the cost sharing all of it right, and the cost sharing is a bit on the local utilities, right, as well, and that's some of the discussion that will happen at this 206 hearing. The economic development that's available, the power data centers isn't going to come to PJM, and I think that's the problem.
And I think that disadvantages states that are pro-business, like Pennsylvania, New Jersey, once you saw that they have a Jersey put in place. Governor Murphy put in a tax abatement, you know, not a ton of money, but wants data centers to come. And you have other states that are huge electricity importers because they shut down all their generation and have leaned on states like Pennsylvania and Ohio that have had gas and gas-fired generation built, and they're supplying. And so there is going to be a lot of rhetoric, but it you know it goes back. I said it, where you stand depends on where you sit, and where you stand, something like that. And your view of what it is, you know, consumer advocates should be advocating for lower prices.
But the reality is you can't have all of these things, right? It eventually comes home to roost, and that's what we're saying. And yes, the political rhetoric is going to go up if this comes. You see people talking about, you know, all sorts of different responses. But I do think PJM has been a functioning market. I do think you'll get, over time, a supply response, and that supply response will be gas. And areas that are prolific gas don't need a lot of pipeline, have space. That means if you want to kind of figure out the Venn diagram on that, that's like Pennsylvania, Ohio.
And to add to Mac's comments, too, I think the other thing is, unfortunately, the focus has been on the rhetoric from some of the consumer advocate groups. In the meantime, you've actually had market participants in PJM say: "Hey, we're looking at accelerating development.
Yeah.
Right? And so it's not, I mean, let's be clear, PJM has two tools to really influence new construction. It's energy prices and capacity. They've given a capacity signal. PJM has done the job that they've done for, you know, 20+ years. So that signal is out there, but you're just not getting as much press from, you know, our peers and some of the private companies that have said: "Hey, we're looking at development. We're going to look at expanding these things.
PJM just put out a release on fast-tracking shovel-ready projects. I mean, so I think this will have a supply response. It's just it can't because these auctions are stacked right on top of each other. That's what Chris was saying. It's very hard to do that in a very quick-
I'm with you. It's a functioning market, right? The compressed timelines make it hard for that price signal to be received in time to, to build a new generation. But this is why and how the capacity markets function. Low supply, send the high price signal, get new build, given the compressed timelines. You know, I think we're in a sweet spot window here of where we can see that continue for a couple of years.
But I think maybe just one other thing. Sorry, this is - I feel like I'm on a soapbox. So, but you know, you get a $270 clear, you get another some dollar clear. I'm not going to put a number on it, okay? That's a $270 or above. And then you get a third one that's there. And if you look at the chart that Chris showed, last eight years, $50 a MW day, RTO. So people who think that you're going to go back to that and have a functioning market, those are price-suppressing, those are price-capping activities that should not be allowed. You should allow the market to work in the capacity auction. I get the point why there should be energy caps in certain situations, but not on the capacity markets.
Go ahead, Shar.
... Hey, Mac. Hey, guys.
Go ahead.
Hey, I think you just kind of touched on my question, but there's obviously a bit of a push-pull between the IPPs and the distribution companies, right? Especially as you're thinking about resource adequacy in PJM. And some of the distribution companies are talking about re-regulating a certain-
Yeah
amount of peaking gas assets, right? Maybe Constellation and Exelon is a little bit too pronounced, the battle between the two of them. But, you know, some of the EDCs have floated maybe working with the IPPs, so around potentially restructuring a state mechanism where you guys can be incentivized to build through a PPA contract, and they can earn on the PPA contract, and ROE, et cetera. Is this sort of a dialogue that you guys are having? Is it so black and white, where it's going to be a battle between the two of you, or can you guys actually strike a middle ground with the EDCs and structure a mechanism similar to the offshore wind markets that we've seen in the East Coast, right? Because they're quasi-regulated and deregulated.
Yeah. Well, yeah, that's a great way to put it, it's quasi-regulated. So first of all, just on load-serving entities or, you know, in the local transmission owner, that we are predominantly into PPL. And Vincent has said, you know, he'd like to explore doing stuff in rate base. You know, if I had a rate base, I'd want to explore doing stuff in rate base. So I have no problem with him doing that, just like consumer advocates want to advocate for consumers, okay? I don't know that that's necessarily the response it should have, because again, I think that PJM is a working, functioning market, okay? That said, there are other things going on. So everybody is familiar with the TEF that's in ERCOT.
Senator Yaw, I believe it is, and Pennsylvania state senator, is putting forth basically a PEF, P-E-F, Pennsylvania Energy Fund. I think it's targeted $5.5 billion-$6.8 billion for 10,000 MW, loans up to 60. We got asked to opine, and you know, my view on that is yes, if that's a mechanism that can work, okay, the finer details matter here. But if it's a mechanism that can work and bring generation, I have no idea how it gets offered in capacity and any of that, but if it gets offered in the regular way, depending upon what strings come with the debt, you know, because it's supposed to be at, I forget, less than 4% type debt, up to 60%, and if that can incentivize, I think that's constructive, okay?
But the devil's in the detail. All right? That's different than rate base, because I don't know how you put that in the capacity, in the auction, in the energy, and the rest of it. So I do think that there'll be creative solution. I think the solutions are going to be... You look at the balance sheets of, I don't, you I can't keep track. What NVIDIA lost, like, $300 million the other day, right? In market cap. But like, billion, I said million, didn't I? Billion. Billion, I can't even contemplate it. But the If you look at their balance sheets and what they're spending on data centers, they don't want to put power plants on their balance sheet, but they can effectively through power purchase agreements. So everybody can have their cake and eat it, too.
T&D companies can build transmission and distribution and put it in rate base. Developers, we might be one, under the right circumstances, can build power plants under PPAs that serve load and also source from the grid. And that's why I keep saying this, this opportunity set is so big for the sector. We should all just figure out how to solve it, as opposed to argue over who gets to solve it before we even talk about what the solution is.
And just lastly, obviously, no data center announcement today, right? So, but just some of the dialogues you guys are having with the hyperscalers. I'm curious, some of the noise you're seeing at FERC and at the state side, and some of the pushbacks with Maryland, Connecticut, is that causing some of these hyperscalers to balk around, to make behind the meter, and instead shift in front of the meter where it's less politically sensitive? Like, how is that dialogue going as this is becoming sort of a federal issue at this point, right? So-
Let Cole answer this, because, you know, I'm out of words, but not really. The thing that I find interesting about that conversation is, and I challenge everybody that starts that conversation with, is this pushing people towards front of the meter? Someone define for me what front of the meter solution is. I'm dead serious about this. We have a solution. We know exactly what it is. We know how we contracted for it. It's behind the meter, direct powered. But people talk about this front of the meter solution. The deal hasn't been designed yet. It hasn't been done yet.
Unless we're talking about doing it the same way that it's been done in the past, which is Loudoun County. I know, pick on, I don't mean to pick on big, but it's the data center alley, right, in Virginia. Plop down a data center, connect it, take the MW if you want them, don't take it, pay your demand charge, et cetera. If that's front of the meter, I don't think that's the solution. But no one said what is front of the meter yet.
And I would say, you know, as I said earlier in the prepared remarks, you know, the what I think we call direct connect solution. You said behind the meter, that's one of many solutions, I think. And it took a long time to figure out the technical on the direct connect solution. So, to your question around, does the ISA give them pause? Perhaps or perhaps not. I'm not going to opine, but I think they're still exploring it. And obviously, the ISA approval and ultimately the FERC technical conference may influence their ultimate decision down the range, but I don't think it's paused them from thinking about the option and seeing kind of how this FERC situation plays out.
To answer that question, by the way, on the FERC Technical 206 conference, we've worked with our colleagues in the IPP space by which to work together to proffer up an agenda and what we think who should be testifying and et cetera, and we're hopeful for that. I would leave you with one last thing. Cole and I and the whole team have been thinking about what that definition of front of the meter is, okay? We're working on that, but it's not the current model. I don't think anybody's come out with a solution as to what is front of the meter yet.
All right, I'll do 63 and then 90.
Hamed Khorsand , BWS.
Yeah.
When do you see the company in two to three years, as far as revenue composition goes? Do you think it's gonna be more of direct connects, or do you see yourself, you know, being current, you know, size and just doing buybacks? How do you see the company in two to three years?
Oh, you want me to go with the hard question?
You go with the hard one.
Yeah.
I'll clean up.
So look, as I said earlier, I think we've got a number of opportunities that sit in front of us that others don't sit. I mean, you know, so we don't have any retail load, so we move up and down with wholesale prices. We actually like that position for the next several years, okay? Because of the things that we think in the market dynamics, and that's why we sold our ERCOT, concentrated in PJM. All right? But we have this AWS contract by which we can accelerate that, and I think what we're looking to do is how do we leverage that and leverage that across our fleet and create more opportunities? And so when I think about where the revenue is going, Terry showed the slide, the Pac-Man type slide, you know, that gets to half. It's like Mrs. Pac-Man colors,
but the gets to half of the revenues coming from contracted streams. I see us heading more in that direction over time. As we try to expand, and as you know, I think Cole said, as we look to get our next data center arrangements, I forget the next-
Power arrangements.
Power arrangements.
Yeah.
So I think that's the direction that we're going in. Now, will we look to add by being part of the solution to build under the right circumstances? Okay, which is the contracted circumstances that we've been talking about. Will we look to add to the portfolio and expand this platform? Sure.
Yeah.
Not by building, I'm saying through acquisitions. Would we look to... You know, we've got we continue to say that we are getting off of coal. There hasn't been, you know, that push sort of died, and if you look at Colstrip, it's making a lot of, you know, money, and it's very important to the economy of Montana and where it is. Okay? But we're looking to reshape the portfolio and continuously do that. Hopefully that's appropriate.
Yeah, Hamed, let me just add to Mac's comments, and first of all, good to see you in person. I think when we think about the contracted margin around the business, our ability to increase that contracted margin ultimately should give us a lower free cash flow yield, right? And we think the value proposition on that is helpful. As we just ramp up the AWS transaction over the next several years, and you heard Cole talk about accelerating that and you know, what that could mean. And then, you know, I think the other thing, and this sort of goes hand in hand with some of the questions earlier, somebody going out there and building a merchant power plant, like a merchant gas plant, like, I don't know if we're going to see those days anymore, right?
What the right solution is, and Mac alluded to this a while ago, is you've got to go back to the way development used to be done in this space about a decade ago. You would go out, you'd get an anchor PPA, something that made sense for a, you know, a decent amount of time, you'd go find the financing around it. It's just today, that finance or that PPA is likely gonna come from somebody who's gonna have, you know, a significant long-term commitment, as opposed to in the past, those are maybe like five to 10-year commitments. We would look to do things like that, either with existing assets that we have today or anything else that we would potentially acquire or build.
But this isn't a new model. You head down, isn't it 95 that goes out and goes through the Meadowlands, right? And the Linden station, or is that on some other. It's on the way to Newark, I forget, it's somewhere out there. That Linden's a cogen plant. This is not a new concept. Building power plants attached, attached to an industrial complex is not a new concept. It's been done, it's just that it hasn't been done recently. It was done, you know. I like to say, when I started as an engineer, I worked at a bunch of the refineries on the Delaware River, and all of them had what they called at the time, a powerhouse. All that was, was a cogen facility, and it was because they wanted speed and access to electricity.
Sounds a lot like the AI boom that's going on right now. Then over time, they became interconnected, or they had a different tariff rate, and I think that's where people are arguing. This is something that happened in the '70s and '80s and started to disappear in the '90s, and it has almost entirely disappeared at this point. There's still cogens left. If you go to the ship channel in Texas, there's still cogens left that are connected-
Oh, yeah.
but also serve load. They also provide steam, which is another thing, but and you don't need that for data center, okay? Because you're trying to do just the opposite, cool. But that's where you're gonna I think you'll see this model with the 10-year PPAs, 20-year PPAs. I mean, that's effectively what we did with the nuclear direct connect AWS deal. Go ahead.
Number 90.
Hi, Greg Orrill, UBS.
Hey, Greg.
Thank you for all the detailed guidance, and my question is similar to this slide, which is: Can you comment on sort of the relative EBITDA of the fossil assets in PJM versus the nuclear AWS?
... assets?
Yeah, let me comment on it without giving you exact specifics, Greg. So a majority of our revenues going forward are from Susquehanna and the AWS transaction on a combined basis. Now, one thing to keep in mind is the capacity clears, if you factor that in, right? We've got some fossil plants in the fleet that are very large assets. We've got, you know, 1,500 MW at one site, we've got 1,700 MW at one site. And so when you think about the capacity revenue component of that, wherever that capacity clear is, that's gonna be a bigger piece of the pie. And so, for example, in 2026, you actually have a little bit of expansion in that fossil portion, but you still have a majority of it coming from Susquehanna and the AWS transaction.
I think, Greg, we still get from the gas fleet, you know, we get the capacity revenues, which when the capacity was at the $50 a MW day, that's when, and you weren't making a lot of energy margin. That's when, going back to what I said earlier, when we emerged, people were saying, well, the gas fleets were zero. Now, you have these plants that are gonna be elevated. You've got CP penalties, so you actually have to invest a little bit more in those. When I say a little bit more, making sure our Chief Fossil Officer, if he's listening, understands what I mean by a little bit more, but that's a joke. But the, but you have to invest in there, but we're also making energy margin on top of that.
It depends on what we see as the energy curve out there. But we do receive, you know, we have 6,000+ MW that clear in the capacity, and only a couple thousand of it’s the new. So...
Thank you.
All right, I'll take two more questions. It'll be number 127 and then 117. And then after that, we'll take a break, get up, get snacks, and probably end the recorded portion of this event.
Okay. There's more questions. We can take a few extra minutes, but...
All right. Thank you. It's Ian Zaffino from Oppenheimer.
Hi, Ian.
How are you guys? Can you maybe touch upon the guidance slide? Are there any assumptions for the RMR in there or any updated thoughts as it relates to that item? Thanks.
I'm going to go first on this, since Mac my first and last one. The answer to your question is yes, there are assumptions in there. So all kidding aside, Ian, we do have assumptions in there for the RMR outcome. Obviously, that's an active settlement discussion that's going on today. I think we mentioned this in the second quarter call. An ALJ was assigned, we've had initial scheduling hearings. We're not going to give the details of what that is. You know, I think we've... You know, there are other RMR contracts that are out there that you can go look at. NRG's Indian River asset was put under an RMR contract this past year, and so that's something.
But we're not going to give sort of specific details about what we have in the model around that.
It's within the range. It's in there.
Yeah.
And I would just add one other thing, that we've asked, and I think both PJM and FERC have acknowledged and appreciate our position on this. I think in understanding that we've asked that that be resolved by the end of this year. Indian River was re-resolved two years in arrears or basically two years in arrears, and we've said we're not going to do that. But we need to have it resolved by the end of this year, and they've been very accommodating to that in these schedules and things of that nature. Do I know if it gets exactly resolved by the end of this year, plus or minus? I don't know. But we've provided a lot of information to them as part of the process.
The reason why it's important to get it resolved is because you can't. It, it's no way to run a railroad, to run right up until the time you're supposed to, you know, set the trains in motion and decide to go. You got to. There's a lot of work, there's labor. We got a couple of hundred employees, we got fuel contracts. We've got to make arrangements for all of that, and so you, you don't just start operating one day later. And so we've asked that it's been. And, and again, like I said, they've been fairly accommodating to us, both PJM and FERC, in, in their understanding of that, that request.
Yeah, thanks. And then, you know, I know you guys have used the term perfect and accelerate several times. You know, what does that exactly mean? I think, we're under the understanding that AWS is basically driving this, but, you know, how do you guys fit into that role as, you know, you move towards 960 ?
Yeah, I'll take that one first, and Mac can clean that one up. Look, so what we mean by perfect and accelerate, simply put, is perfect just means, as you saw in one of the financial impact slide, there's an option out at the 480 off-ramp, right? And so there's some things that we are helping, but ultimately, Amazon's driving to ensure that the campus can get to 960 MW. That included a lot of the development milestones that we talked about. We obviously spent a lot of time over the last three years building the campus, and yes, we sold it to Amazon and AWS, but we have a lot of knowledge of the infrastructure, the power infrastructure that's going to be required for the eventual build-out.
And so we are supporting from a, you know, knowledge base, to help them accelerate, and perfect. So perfect is get to the 960, accelerate is do that on a time frame quicker than the 120 minimum ramp schedule they've committed to.
I think, Ian, that's the project management aspect of it, and it's the most valid. But if I convert it to sort of what I think would be most relevant for a set of investors is, you look at AWS, we've got some cash flow coming off of it. The contract's already started, okay? But it's effectively a non-cash flowing asset, all right? And so what people do is they look at these projections out here. Now, there is cash flows in these projections as we described, but what people are looking out there is, they're putting cash flow projections based off the slide that Cole said we didn't change, just so we didn't fool anybody on the reference price....
And they're putting an uplift on that, and so they're thinking, okay, well, how much of the future cash flows are going to get out there? And people are risk adjusting that. I don't know that people are fully valuing the 960, so they're risk adjusting it. So we want to perfect that people put the 960 in, and then if you do that, it's on a ramp schedule that goes out there. If you can accelerate that ramp schedule, that's worth a lot of DCF value, and I think that's how people are valuing. It's what I hear a lot of feedback from investors are. They look at the base sort of operations and think about either multiples or free cash flow yield, et cetera, and then they're doing a DCF with respect to the AWS contracts.
And so we want to perfect the 960 in people's minds, which we don't control, as Cole said, but we're doing everything we can, you know, whether it be at the supervisor or township meetings, to make sure that things go through working with AWS to help them do that. And then, you know, there are 15-16 buildings tenanted by 2028 that would bring it earlier than what we put on those slides. So that's how I convert it to the financial metrics. Does that make sense?
Thank you. Thanks for squeezing me in. Durgesh Chopra, Evercore ISI. Maybe just, Mac, just your thoughts on construction timeline of a gas-fired power plant. So your peers talk about multiple years, and I hear you and your peers kind of looking at this PJM price signal, but not wanting to jump in. Just how do you marry that to how do you make sure that, you know, you have the equipment right in time, you have the permits right in time, so you can timely respond to those price signals?
The honest answer is that we've explored some of this. Cole's worked on this with Dale, our Chief Fossil Officer, and spoken with turbine manufacturers. There's not a lot of turbines available before a 2028 delivery, okay? And the old model of putting a reservation down at, like, 10% or 20% and having a spot in line and, you know, for the next turbine, et cetera, and then being able to sell that in a secondary market, that doesn't exist. I mean, GE Vernova is saying you gotta pay for it all to get a spot in line, and most of these turbines are going. Who knows? That's a negotiation, but a lot of these turbines are going to AMEA, if you will, Middle East, for the most part.
Now, will someone in the Middle East decide that, you know, crude's come down, maybe we'll just burn more crude, and we'll sell those turbines back to somebody else, you know, at a premium? Who knows? But I mean, it, you know, we have a lot of sites that we could redevelop that already have water and interconnection agreements, et cetera, but the supply chain's an issue. By the way, I was speaking with somebody else, and there's one thing that I think is very interesting is that people don't realize that the supply chain is getting a little bit clogged up, if you will, from a financing standpoint, both in thermal generation, right? No one's really financed a CCGT for years, right? So you got to get that unclogged.
Then money flows that used to go into renewables, by the way, for wind and solar, for some of these things that are in the queue. Everybody knows that re-rated about this time last year, right? When a company came out and said the renewables sort of cachet is off or whatever it's called, the patina. And so money flows have stopped there. Now, there are people still building. I mean, you got relationships between, like, one of the big renewable developers and Microsoft to go do 10 GW and things of that nature, but that system is starving for capital, right? I mean, you all write on this, and so it's sort of another additive of why it's good to have existing generation on the ground in the right spots.
Durgesh, maybe to add to Mac's comments, he talks about right engines, but the other part of this that you've got to sort of take into consideration is what about the transmission infrastructure?
Yeah.
Right? Because the demand on that is sort of across the board. It is not just, you know, the T&D guys are out there, you know, building more, you know, building more long-distance lines, putting more transmission in place. And I think when you talk to OEMs on the transmission side, that actually is a little bit of a longer pole in the tent than even the engines. Now, on the engines, whether it's a CCGT or a combustion turbine, I think there is a slight difference. You can find some combustion turbines and aero derivatives, you know, sooner than, like, a large-scale CCGT. But I think the real holdback is ultimately again, on the transmission side.
But look, I think these things will solve themselves over time. It just takes a little bit of time, I guess. As I was sitting here listening to that and listening to myself, you know, I was at Exelon, which is now the Constellation part, in 2002, and I was listening to a presentation by, I won't name the market consultant that came in and said, "You know, SO2 rates are gonna go like this, through the roof," right? But this is the Clean Air Act Amendment, final trading SO2 allowances, and this. They came in and said, "You gotta put scrubbers." This is actually on Keystone Conemaugh. I can't get away from that thing. And I love the guys at Keystone Conemaugh, but it's just, it seems like everywhere I go, we own a piece of Keystone Conemaugh.
But put scrubbers on it. You're gonna need to put scrubbers. Ultimately, we did on one of the units. I can't remember, I think it was a Keystone unit, et cetera, but not all four and, you know. And the reason, the rationale was, is that the supply chain, you're not gonna find enough welders, you're not gonna find enough steel, you're not gonna find enough engineers to design all this because everybody's gonna be doing it, so you better go fast. And this feels a little bit like that. It'll solve itself over time.
... Time for a couple more, you think? Maybe one or two more.
Sure. 114, I think has been-
No, I-
Okay, good.
Let me do one 31 actually, back there.
Hi. Thanks, Paul Zimbardo. I'm used to being last, so it's okay. Just one follow-up on just the M&A side of the conversation, like, kind of a two-part question. Like, first is you have a lot of opportunity set, the AWS ramping the other sites, like, why entertain M&A at all? It's like kind of a broad question. And then the second, like, if you do entertain it, you've talked about, like, you exit Texas, really like PJM. Would you want to stay in PJM, or are your ambitions broader?
Paul, I appreciate you asking that because I may not have been clear or precise with my prior answer with respect to M&A. We don't comment on M&A. What I was talking about was buying and selling assets and looking to add opportunities or to hive off opportunities. We make zero comment on M&A. So if I wasn't clear on that, I'm not trying to... Well, I am trying to shut you down because we no comment.
Okay. No, that, that's clear. Thank you.
Looks like Angie may want to sneak in one more.
Just one other thing. And again, I think I made this point before. You remember you mentioned that your target leverage is three and a half times-
Mm-hmm.
Net debt to EBITDA. You wish you had bought more stock at a lower price.
Yeah.
You have, you know what your EBITDA trajectory is. That would imply, again, using your numbers, that you have probably about $3 billion more of buyback capacity versus what you had just announced. Again, using your own numbers and the leverage target. So is it just you're pacing yourself? Again, just explain to us why wouldn't you actually have a much bigger buyback? And yes, probably it's coming, but just trying to understand if that's basically the way you think about it.
That's the way we think about it. And as I said, and yes, we're pacing ourselves. And third, yes, we've set a target of three and a half times. I think your math is probably. I'm sure your math is right. It depends on what you use and those ranges, et cetera. And that's why I think we described the balance sheet as a strategic asset that we can use. I think that, you know, at the second quarter, you know, there's time and place to talk about and to do things, okay? In the second quarter, people said, "Well, why didn't you up your share repurchase program at that point in time?" It's like, well, first of all, that earnings call was in the middle of August.
It was, you know, not this well attended, which we appreciate everybody being here. And second of all, we were just, you know, we had just rushed to do a whole bunch of other things, tenders, bilaterals, et cetera. And so today, we're up in that share repurchase program, and I think we've shown the discipline to look at that, to manage our balance sheet, to view it as a strategic asset. And if we, you know, and we do say, and I think Terry said, highest and best use is a share repurchase program. So I would tell you that all things are on the table. And, you know, we would have to completely, just be clear on what we'd have to do.
Terry often blames me because I was there on exit for the exit financing package. For those of you who bought the debt, we appreciate you, but you also, you know, got a good package. You know, the bonds are trading at, you know, I forget what they're trading, $107, $108. They have a make-whole package, so there's costs associated with doing things, associated with changing the balance sheet. So we're just weighing all those different options. It's not as though we're not thinking about them, but it is a time and place, and we're contemplating it, Angie. I hope that answered or was insufficiently clear.
We'll sneak in one last one with Thomas Meric.
Thanks, thanks for the time.
Yeah.
I want to come back to kind of commercial IP around data center deals, and just think about broader market in the context of a bottleneck, knowledge bottleneck. Whether it's from the data center, understanding power or an IPP, understanding data center economics and commercial realities. Does the solution provider, the entity, that kind of, you know, quote, unquote, "win" for everything about it, manifest itself in speed to market or price? Just, do you think it's some combination of both, obviously, but maybe just any thought from it?
Yeah, I mean, I think ultimately, it's a combination. It's an all-in package, right? I mean, obviously, go back to the slide that we talked about. It's about the speed to market. It's about how much can you get, right? So if you can get 200 MW or you can get 1 GW, that matters to folks because they're putting in hundreds of millions of dollars of fiber infrastructure. And so there's huge efficiencies to doing it all at one site versus doing it piecemeal in smaller blocks. Obviously, price matters. I think it's the number and whether it's fixed or floating and all that, but also how long and giving them that certainty. Because it's not. This isn't a 5-10-year PPA they look at it as.
They look at this as, you know, just use Susquehanna as the example, life of the asset, right? 2040s, 2060s. And so I really think it's all of the above to kind of go into it.
And maybe just to build on what Cole said, it's. I think the power side needs to get smarter about data centers, and I think the data center side needs to get smarter about power. It's that simple. That's why I was saying, while we're at an inflection point, trying to solve all this new demand, okay, we don't understand the demand, but the demand doesn't understand the supply either, okay? So take the five nines that I was talking about, and take the-...
It's easy to understand when you say, "Reserve me 40 MW," but I only take 20 MW, that's an expensive solution, right? Because then that 20 MW is just held in reserve, and it's waste, okay? But take the five nines, for example, and take business interruption. Someone once told me that the reason why what I'm saying, or about to say right now, doesn't matter, is because converting electrons to data is the most profitable business in the world, okay? Yes, that's a great way to look at it, but when the data center lights stay on, and potentially other lights go out, that's not a good look, okay? So what would your business interruption costs look like from the demand side?
And I actually think this is where there's a white knight opportunity from the data centers to say: Look, we'll be willing to either lean on our diesels, right? That are our primary source of backup in most data centers. Lean on our diesels or cut, right? A lot of this stuff is being used for learning AI. Now, I wouldn't profess to even know. I don't know what it means, but like that, that's it, don't peel the onion one more layer. But a lot of those algorithms are written to where they run continuously. Well, why can't you run them batch, or why can't you shut them down for 24 hours? If you could do that and then put the load back on the grid, you put the MW back on the grid, you inherently have a DR response that is highly worthy.
Highly worthy?
Highly-
I don't know, whatever, but it's worth a lot of value to the grid, and you look... Yeah, I know. We share. As I said, my English is a second language, so English being the primary one, too. But you get my point, though, right? I think that people from the data side are gonna have to get more educated about power, and power is gonna have to get more educated about data. I would say that we have a head start on that and have learned a ton, but there's more to go, and you're gonna. That's why when Cole says we have one solution, who knows? Our solution may modify over time, right?
We may have another solution at another site that looks totally different, but I think it's gonna have to fit in with: how do you think about these things? I did talk to one person off the record, there was a data center person, and they said, "Five nines is not necessary," and by the way, it's a falsity. It's false, because there's not a grid in the system that provides five nines of reliability so if someone's selling you that, I don't know how they'd back it up, but... so it's very interesting that we're at this inflection point. I think it'll be talked about for the next several years.
All right, well, I think that-
Well, yeah, just, thank you. Appreciate everybody's-
Mac.
Yeah. No, it's okay. But appreciate everybody's interest in Talen. This is a tremendous turnout. I know it's late in the afternoon, and many of you were over at the Barclays conference and doing that. So we appreciate your interest, and you know, we do think we're uniquely positioned, and we're excited about the future of Talen and powering the future. Thank you.
All right, team. Feel free to grab snacks and hang out if you'd like, and also, there are swag bags outside, so grab those, too, if you want. Thanks again for coming.
Good seeing you.
Good seeing you.