← Currents
Currents
Ep355: How Data Centers Are Reshaping Project Finance
00:26:33
Ralph Cho, co‑chief executive officer and founding member of Apterra Infrastructure Finance, joins host Todd Alexander to discuss the rapid growth of data center financing, evolving debt market structures and broader trends in project finance. Plus, a...
Loading summary
Transcript44 lines
- [00:00]
A
Foreign. Norton Rose Fulbright podcast. Today we're recording with Ralph Cho, co chief executive officer and founding member of Aptera. Aptera is an infrastructure finance platform that works in alliance with Apollo Asset Management. Ralph's joining us today to discuss the financing of data centers, as well as to give us a general update on the project finance, debt markets more generally, and when I first started recording this podcast about eight years ago, Ralph was the first non Norton Rose Fulbright guest that I asked to record with me. And today I specifically asked Ralph to record with me because it's going to be my last time hosting Currents. So Ralph, thanks for being there. Full circle for me after today. My partner Jim Berger will be taking the reins.
- [00:53]
B
I can't believe we've come full circle. And I'm super honored. I am like your first and your last.
- [01:00]
A
Yeah, yeah, I can make some jokes about that, but I got to be careful.
- [01:06]
B
My mind is probably not appropriate, but sure. Let's talk about our favorite topic, digital financings in the market.
- [01:14]
A
Yeah, yeah. So let's not get too sentimental or do off color. So getting down to business here, just to give people a flavor, how big is the data center market and how does that compare with the rest of the projects market?
- [01:28]
B
Well, let's take a snapshot of what we've seen in the first quarter of 2026 today. I can tell you, looking at the latest numbers, roughly we saw about 70 billion of project financings and infrastructure hit as of March 31st. And I would tell you, looking at the digital market, it's about 35% of that 70 billion today, probably like around, you know, 26, $27 billion at least. So, you know, if you kind of annualize that, I think we're headed towards over 100 billion just for this year,
- [02:00]
A
which to me is just mind blowing to think that something that's fairly new to the market, I mean, this has only been in the last whatever, two years or something like that, that the market has had to absorb all of these additional financings. And it seems at times like the demand from these data centers for debt financing is nearly insatiable. Where have the funds come to do this since the project market was already very strong? And where do you think the money's going to come to finances, data centers going forward over the next six to 12 months? And I guess that's a really good question for Aptera because you, unlike many of the other people financing these types of projects, are more agnostic in terms of where you can source funds from. So maybe you have better insight than others into where pockets of cash that haven't been tapped or that are tapped but still have some dry powder, if there is anybody that still has some dry powder, can still take additional exposure to the data center market.
- [03:09]
B
I agree with you. I mean there's so many of these data center financings that we've seen in the market. And I mean this statement might be a little bit controversial, but at some point, Todd, they all start to look the same, you know, and they're huge. And you wonder like who's taking down all this paper for the most part, where all the volumes and activity that you're seeing have been really around construction financing for a lot for some of these hyperscale, you know, tenants. And these financings can be as small as like, you know, whatever $500 million. I've seen them go all the way up to like multiple billion dollars. So they can be super large. And now with the whole onslaught of these data centers bringing their own power, the financing sizes are even getting even larger because it has to account for the power plant as well. But in general, we have seen the commercial bank market really super active and they're probably, they're probably the most efficient lender around it especially because there's some construction and delay draw element to this. And we've seen commercial banks finance them for hyperscale tenants. And it doesn't matter if you're a triple A rated entity or a triple B minus entity. The pricing has held pretty steady around 225 over SOFR, maybe call it plus or minus 25 basis points. And what I would say is this more creative capital is coming in. Obviously you have guys like us who are more private credit. We've certainly seen other type of private credit and other asset managers come into the fold and bring hopefully try to bring lower cost of capital or ones at least that can compete with the banks. Not as many, but maybe they are been trying to be a little bit more aggressive to justify premium yield. Insurance companies certainly have been coming in, especially if you can get these to either be rated investment grade. Some of the insurance companies might get an implied investment grade rating. So that brings institutional capital into the fold. For the most part. That's where the majority of capital has been coming in to finance like these, Greenfield, obviously once they hit operations and hit cod, you can go other places. You can go to the ABS market. We've seen some of the financings done there. We've started to see more and more 144A lender institutional loan market come in and take out some of the financings. But you know, this is all obviously relatively new.
- [05:29]
A
You think the commercial banks are going to have to either get very short tenors or have step ups very quickly after construction finishes. Because at some point the bank market's only so big and there's billions and billions of dollars of these being financed. I'm sure they would love, I mean they don't have to, but they would love to get to be able to recycle their money quickly and put pressure on sponsors to take them out with longer term financing once they got construction completed.
- [05:56]
B
That's the story, right Todd? I mean that's the story that they sell internally to their credit. And the way most of these loans are all structured is like a three plus one, plus one type of tenor. And what that means is they get the three years to build the asset and then there's a plus one, plus one. So they, the borrower can add at their option two more years of extension on tenor and that gives them some time to get the asset refinanced. I haven't, I mean I haven't seen anything that veers really from the structure that they put in today where basically the borrower has to pay an extension fee right, to get those additional years. But given the value creation of what these assets are worth pre during construction and then post construction and there is definitely a jump in the valuation, most borrowers are incented to take out the financing. Because I have to tell you, if you look at some of the structures that the ABS market has been refinancing, these operating assets, they jump from an ltc, which is a loan to cost metric, to an ltv. And, and in most cases the borrower is able to get all of their capital out plus some kind of a dividend at the refi. And so you have a lot of incentive borrowers to exit into a permanent financing structure because of that favorable refinancing terms.
- [07:17]
A
How do you think the fact that some of these guys, or maybe more than some, maybe many, are bringing their own power? I was saw today that there was a big announcement of a huge multi gigawatt project announced in Texas where they're going to bring their own power is that change either do you think the types of banks that will do the deal since now you're both taking power risk and data center risk, or do you think that the fact that you have those combined together maybe some people will think is a good thing because now you have control over more vertical integration and more control?
- [07:55]
B
Yeah, I mean Look, I have mixed feelings on this personally. I mean, it totally makes sense that the data center owners are bringing their own power. It creates a lot of pressure on the grids to be able to provide the power to the data center. And I think it unfairly distributes the cost of new generation onto consumers. And so it makes sense that if you are a data center owner, you should bring your own power. So I get that. And in some cases, like I said, that's a positive thing. I'll tell you, when you want to finance a power plant on its own that's fully contracted to an investment grade off taker, no problem. People do that all the time. If you want to finance a data center construction with a hyperscale tenant, also no problem. We've seen obviously a lot billions of dollars financed on both sides. Putting them together brings a whole host of new risks and issues. And project finance lenders don't take project on project risk. They don't do that. But here what we're starting to see, and there's more of these types of deals coming into the pipeline. You know, we see lenders willing to take, you know, project and project risk. And so what I mean by that is that you have a data center and you're already taking the operating risk on the data center, albeit it's all high grade, you are still taking risk. And now you're taking risk that you have a captive power plant that only can provide power to you. And if for whatever reason there's issues with that power plant, you're stranded. That means that you have a data center with no power and probably a tenant that will walk away. Because I'm sure there's some contractual provision that the hyperscale tenant can walk away if they're not getting any power to the power plant, not power to the data center. Then conversely, that power plant can't sell any power to anybody but your data center. And so that's what creates the issue. Now I think it's all structural and doable and executable and I'm sure at the end of the day it'll all work out right? These contracts are highly negotiated, obviously, and people are very, very familiar with each of the risks individually for both asset classes. However, then the whole question, at least in my mind comes is what is the premium that you should be paid to basically take on all this extra headache and risk analysis? I've seen one transaction closed to date. It's like a multiple billion dollar deal that actually had a power plant and a data center and just for people in the audience to know, a data center hyperscale tenant by itself usually, like I said, clears around 225 power plants if they have a utility scale contract. Also probably clears in that today in that 225 area. So in my mind, I think generally you would expect there to be some kind of a premium, you know, than the two to finance both of these together. What is that premium? That's the million dollar question for me. I would imagine that you need to have enough premium of a yield to justify taking on all this additional risk. What is that, 50, 75, 100 basis points? You know, I think any of these make sense probably on the wider end for me. But you know, obviously everyone has their own view. The financing that got done recently had a 25 basis points premium, super tight. And so that means people are willing to take on all that extra risk for 25 basis points versus doing a data center clean without the headache and burden of taking on all that extra risk of being captive to one power plant. Does that make sense? I don't know. But I'll tell you, I think that financing that closed recently, the market did super well. And so I'll tell you there clearly is a market and appetite for that at that 250 level.
- [11:43]
A
Kind of comparing and contrasting the data center market, which is still evolving and is a little bit different from the traditional power market, which is the biggest area of projects otherwise being done. I don't know if data center is going to be bigger than power pretty soon. You've talked about loan to cost. What's the typical leverage you're seeing? Tenor and corporate recourse, things that are slightly different than the data center market versus the power market.
- [12:12]
B
Look, well, let's talk about the data center market. And yes, I do believe that the financing volumes in data centers could surpass power even though they go hand in hand. You know what I mean? Just again, I'm looking at first quarter data in 26 and renewables alone probably represented about, call it 30% of volumes. And you know, you have data centers representing like 35. So already it's surpassing it. In the first quarter alone. Typical leverage levels for construction of data centers has been, let's call IT market around 80%. 80 to 80% of LTC, which is loan to cost, we've seen it go, go as high as 95% for certain borrowers out there that have strong relationships with their lenders. But that's kind of like the range, let's say 85% ish. Right. Once these assets get into operations like I mentioned before, they get valued on an LTV basis and those probably can go to like 65% of LTV. And that's definitely much higher than LTC. Right. So that's where I say that the borrowers are motivated to refinance these assets. Tenor three plus one plus one. So call it five years up to five years if the borrower needs that. And then pricing, like I said at 225 over, you know, from Aptera side we've been able to be a little bit more creative and offer borrowers financing at Holdco level. Holdco financings can get done. Call it at a higher LTC, like call it 95 to 100% of LTC. Right. And we price those at let's call it 375 to 400 over and we'll give borrower, you know, call it five years of tenant on something like that. That's generally where we've been playing in. I would also kind of mention that we've seen edge data centers. These are data centers that are much, much smaller and maybe more in like not non tier 1 locations and closer to the user that some of these developers might build. And they'll still sign up with hyperscale, you know, investment grade counterparties, but they're just smaller. And so wolf financy is very similar to a hyperscale data center. And we'll go to like 80, 85% of LTC. But because of the size of these being smaller and in more remote locations, I call random locations, the pricing, we've been executing them anywhere from like call it 250 to 350 over SOFR. So there's definitely a premium on doing something like that. I kind of liken it to the renewables business of doing like a utility scale solar project versus like a distributed generation asset. You're going to have a little bit of premium just because of the size, if that makes sense.
- [14:54]
A
How do you think about the refinancing risk? Because these projects, I don't know, the average size now is like you know, $3 billion or something like that. I mean they're just enormous.
- [15:02]
B
Yeah.
- [15:02]
A
And even if you do a three plus one plus one, okay, so you get out two years of operation. The reason they haven't refi ujet is because there's a problem. Of course, you know, when you're doing the underwriting, how do you assess the refi risk because you're still holding it until you're taken out.
- [15:21]
B
I mean, look, keep in mind this at the end of the day, we're sizing these on cash flows, right? And so we're sizing them, like I said, their metric is of ltc, Right. So you're inside the costs and equity has skin in the game and we look at the hyperscale contract. So in the worst case, Todd, even if you don't get refinanced as a lender, I can sit there and post maturity, start sweeping cash. And because I sized it against the contract tenor and the terms in theory and hopefully in practice, I'll get my money back eventually because I'm getting paid by an investment grade tenant. So that's where the comfort comes in now because if you've seen where these assets are being valued, I feel comfortable that I'm in the money. I know that I can sell this asset and get more than my debt leverage back. And in the worst case, I sweep. That's where the comfort comes from.
- [16:16]
A
So in practice, it's basically the same principles you use on any other contracted asset that you're doing project financing on the market has the same basically discipline.
- [16:26]
B
Yes, I agree totally.
- [16:29]
A
And how about on the whole code you were talking about there? At least on most of the power deals, the whole co usually more like a 50, 50 split of the balance of the equity, they still want some skin in the game, what I've seen.
- [16:42]
B
But yeah, I mean, like I said, I'll go to 95 to 100% of LTC at the whole CO level. And so sometimes there isn't as much skin in the game in that case. But again, the same theory applies because we're looking at the underlying cash flows and we know, especially for some of these rich contracts, I know that there is enough distributions after paying the OPCO that there's still cash flowing upstairs for me to take the rest and pay the holdco lenders off. Right. So again, it's all cash flow based. We're not really taking a view on, on the valuation of the asset to sell, even though if we did, we're still in the money on the holdco we can sweep out with cash flow. I think it's also fair to mention that we've encountered in the market sometimes some of these data centers have such lucrative contracts out there that if you sized it out using like, you know, coverage ratios, like, you know, whatever 1.15, 1.2 times, you can size out debt that exceeds the total construction costs. So let's call it 105% of LTC. Right, 110% of LTC. If you had no cap on LTC, right. You could in theory a developer could get cash back and have no money in there. So in principle, fundamentally I think many lenders would have a problem with this. Right? Because nobody wants to finance a construction project and start giving money back to the developer that they have no skin in the game. I think we personally would have a difficult time giving cash back. I have heard in the market that there are some lenders that are, if they're willing to be paid enough, they might be willing to do this especially for proven developers. Again, I haven't seen anything have heard about this. We've been told that we've lost some transactions because you know, the developer was able to procure a high ltc, you know, financing. Again I, I just want to throw that out there that if you have a lucrative contract that that could potentially be an option. But I'm sure it's not like cheap.
- [18:44]
A
This reminds me of years ago there were those very high priced, what turned out to be high priced PPAs because the cost of equipment in renewals kept declining so rapidly that there were a lot of these PPAs are expensive and if you use debt sizing metrics that were just based off of coverage ratios based off P50 you could have negative equity. But pretty soon the lenders figured out that was what was going on and said no, no, no, no, got to have some minimum equity here.
- [19:11]
B
There is such euphoria around this space Todd. And there's so much capital that's coming in, it's crazy. Like deals, multibillion dollar deals like oversubscribe and lenders are holding huge amounts and so we've definitely get approached by a banks that have huge data center books and they want to unload and they want to sell to us in a package deal. So we certainly talk to lenders about that. I think the key is you have to keep, if I'm a syndicator, I have to kind of keep finding new sources of banks that haven't been as exposed to data centers and wants to jump in. And you got to keep kind of like laying off the risk to smaller retail banks. I mean that's one way to kind of manage your book. So you're either doing portfolio sell downs or you're managing retail, retail sell downs or you hope that your borrower is going to refinance you especially when the assets hit cod. I think that's the way that these capital is going to continue to recycle.
- [20:10]
A
So let's talk about that as we get kind of towards the end of the conversation here. How much of the data center demand is ultimately going to crowd out the demand for other types of project market? Or if not necessarily crowd them out, but cause a tightening of terms if someone was able to do a deal at a fully contracted 25 year bus bar BPA or something like that. So for Plus 1375, is that now going to go at 150 or 175 because there's just not enough capital out there or are people going to tighten terms or are you not really feeling that pinch now and there's enough global capital still available to service what's there?
- [20:52]
B
You're talking about like supply and demand, right?
- [20:54]
A
Yeah, supply and demand.
- [20:56]
B
If there's so much supply and there's only limited demand from the banks. Right. At some point you think pricing has to adjust. Totally agree with you. And initially when we started Aptera three years ago, that was like our thesis. We said, hey, the pricing has to at some point lift three and a half years later. Haven't seen that happen. In fact, the opposite. It's been super stable at 225 over.
- [21:22]
A
All right, last question for you and I guess last question for me.
- [21:27]
B
This is historic, Todd, make it a good one.
- [21:30]
A
But it won't be that good of one. It's just switching off because we have you on every year to do cost of capital, which you did in January. So we're kind of at the middle of the year now.
- [21:40]
B
Yeah, I love, I love that.
- [21:42]
A
Overall, what do you see in terms of trends in the project debt market, period, Putting aside data centers because we've kind of gone into depth already there. Yeah. And anything that's surprised you that's either better than or worse than what you expected when you recorded with us in January.
- [22:02]
B
Oh, wow. Well, let's first talk about the trends that I see in the market. Obviously we're still doing a lot of renewables and at least for our capital, the distributed generation community, solar actually works out very well relative to the utility scale. And that's just because of the premium in the market and we think that that's a good risk return for us. So we've been doing a lot of that. But I expect at some point that renewables is going to start to slow down. Right. There's only so many kind of safe harbor assets out there. Maybe some of the larger platforms have Safe harbor through 2028, but I think just my expectation is that they will market, the volumes will slow down. Although what's interesting is to replace that volume. I think we're starting to see more and more gas assets. And so whether it's a new construction, which is a little bit more rare, or acquisition or refinancings, that volume of activity has started to pick up. Because if you own a gas asset today, it's worth money and you can get cheaper financing or dividend recaps relative to where we were, call it three, four, five years ago. And so we're starting to definitely see a pickup in gas. And I'll even tell you we're even getting asked to look at some coal fired assets, which is really interesting. The only other trend that's kind of worth noting is I know you said aside from digital, but the digital market continues to evolve, right? And so aside from the hyperscale tenants, I mean we're now seeing creative counterparties that don't necessarily have ratings, call it like anthropic or OpenAI. And so people are trying to figure out how to do those financings that make sense. And by the way, to make things even complicated, they're bringing their own power. So these financings are starting to look very, very large, call it 10, 15, $20 billion. And so I think it'll be very interesting to see how we can execute these financings. And you're going to need large capital providers to fill the gap to do those kind of deals. So that's kind of what I see in the world of like, of trends today. What was the second part of your question about that?
- [24:19]
A
Was it just basically anything that surprised you on the upside surprised you on the downside compared to what you had reported in January?
- [24:25]
B
Surprising. It's kind of what I said to you before. The market is so aggressive and there's so much liquidity and I have a hard time even differentiating all of these different data center deals because they all look very, very similar. I'm just surprised at how the market continues to print these at super tight pricing. And more importantly, when you people introduce risk like behind the meter and you know that, that a data center has to guarantee to their hyperscale tenant what they call five nines, that means that power plant has to operate at 99.999% of the time. And sure you can have all the redundancy and the battery backup and everything that you need to give confidence. And I'm sure it'll be fine. But honestly like you know, 25 basis points of a premium, Come on. I'd rather like why, why is it, it's not even worth it. Like I would rather go and like just do a clean deal all day long. So that's what. I'm very, very surprised that the market is willing to do that. That's great. That's great for borrowers, you know, that's great for developers. Great for borrowers. And as long as that liquidity is there, you know, if I were them, I would certainly keep tapping all the aggressive capital as I could.
- [25:33]
A
All right, well, with that, I'll sign off and. Sounds like we'd both be busy at least through the rest of the year, given the growth in the market.
- [25:40]
B
I hear you, Todd. And you're. Now you're retiring. You're handing the reins over, man. Another. Another era behind us.
- [25:48]
A
All things must pass.
- [25:50]
B
All right, Very good. Well, thank you. Thank you, Todd, for having me as your final. Your final guest. I love it.
- [25:56]
A
My pleasure.
- [25:57]
B
All right. Thanks, man.
- [25:58]
A
All right. All the best. You can find us online at www.comprojectfinance law or send us an email at currentsordonrosefulbright.com Please rate, review and subscribe on Apple Podcasts, Spotify, or your preferred podcast app. Our show today was produced by Emily Rogers.
- [26:18]
B
Stay ahead of the Currents.