416. The Bond Market Bites Back: Treasury Intervention, AI’s $3 Trillion Bet, Office Loans Below Breakeven, & Forced CRE Sales

0:05Welcome to the Trepwire podcast, the show where commercial real estate meets data and insights. This is our Week in Review for the week ending August 21, 2026. I'm Haley Keen with TREP, a data, modeling, and analytics firm for the CMBS, commercial real estate, and CLO markets. I'm with Lonnie Hendry, Chief Product Officer, and Stephen Bushbaum, Head of Applied Research and Analytics. This week, the bond market and Washington's response to it moved to the center of the conversation. The 30-year Treasury yield briefly climbed above 5.3%, its highest level since 2007, before Treasury Secretary Scott Besant stepped in with plans to at least double the size of certain long-term Treasury buybacks.

0:51That announcement sent the 30-year yield lower by roughly 10 basis points. At the same time, the minutes from the Fed's July meeting showed that concern over inflation was broader than the three formal dissents suggested. More officials favored raising rates, and many indicated that additional tightening could be necessary if inflation fails to improve. The economic data offered a mixed picture. Manufacturing strengthened, but housing starts and pending home sales declined sharply. AI also continues to appear everywhere, from massive data center commitments and unusual financing structures, to growing concerns about power, regulation, and whether all this investment is contributing to higher borrowing costs.

1:36Later in today's show, in our Digging Through the Data segment, we'll examine performing securitized office loans with debt service coverage ratios below 1.0x. We'll look at why many of these underperforming loans are backed by surprisingly well-occupied buildings, how free rent periods and concessions may be suppressing cash flow, and why the upcoming maturity schedule could turn today's operating weakness into tomorrow's refinancing challenge. But first, Stephen, how should we reconcile a Fed that is still discussing rate hikes with a Treasury department that appears increasingly uncomfortable with rising long-term yields?

2:16I think that the short answer is that the Treasury announcement changes the near-term market dynamics, but it doesn't really change the longer-term problem. And in fact, what's really hilarious in a way, I mean, in a dark way about the Treasury move was they announced the buyback. The long end of the yield curve rallied. So yields came down from their touch of almost 5.3%. And they came down about 10 basis points, which is a huge move for the long end of the curve like that. So that was great. The short end of the curve didn't really shift. But then Thursday morning, six-tenths of that move completely reversed.

2:56So in basically 24 hours, the market said, all right, Prince, we passed$40 trillion in national debt because of these buybacks, or at least a part that helped push us over the edge. It wasn't going to take much, but that got us there. And essentially, we're back to where we started almost. So clearly, the market is in charge. And the really interesting part here is that Kevin Warsh's comment about play the ball, not the referee, just seems to be seasoning worse and worse. I mean, it's rough to watch this play out in real time, but this place is so much more importance now on the Jackson Hole Forum.

3:34I mean, his comments there will be critically important because markets have had a bit of a tantrum, but also, I mean, they've taken the lead for sure. and I don't know how sustainable this is. So let me get back to the buybacks. I wanna be clear about what the treasury did here. This wasn't quantitative easing and it wasn't the treasury eliminating debt. The program is designed to buy older, less liquid securities and improve trading conditions. And the timing matters here. After the 30-year yield reached approximately 5.34%, that was obviously the pain point. The 5.3 % threshold seems to be the magic number.

4:08It was a 19-year high at that 5.34 % And the Treasury just came out of left field and said, we're going to double these long-term buybacks from about$2 billion to$4 billion, at least$4 billion per operation. So it was good news. 30-year yield rallied about 10 basis points. The dollar weakened and rate-sensitive assets received some relief. But that$4 billion is an absolute drop, absolute drop in the bucket. I mean, the Treasury market is so much larger than this program. The larger story here is that really the market needs an enormous amount of capital. We talked about this last week. The federal government is running large deficits.

4:46Countries are increasing defense spending. Companies are reshoring subchains. And the AI buildout is a massive source of capital demand. And so these AI companies are no longer acting like capital-like software businesses. They're acting more like infrastructure and power plays. And so that sort of scale requires massive capital. And that's not going to let up anytime soon. So you combine that with the concerns about longer term inflation, and it's a rough backdrop right now. But we got those Fed minutes on Wednesday, Lonnie, and they mentioned that we had three descents, but it sounded like there was really more support for a rate hike than just those three descents.

5:24So do you think we're going to get a rate hike here at the next meeting? I think the consensus is going towards the next move being a rate hike. I don't know that we're going to get a rate hike at the next meeting. I think the minutes that we saw this week, obviously, were before we got the softer inflation print, before you got the jobs report, the most recent one. So I think maybe those factors come in and say, look, we don't need to move rates up yet. But I definitely think there's certainly more agreement than the three dissenters would have you believe. And so, I mean, you know, obviously the three dissents represented the largest same direction break with the Fed policy decision since 2016.

6:07So just having three people dissent, pretty big news. But like you said, if you go through the notes or the meeting minutes, there were certainly more people that are on that train of thought that, you know, we probably should be raising rates before we do anything else. I think that was part of the narrative was just that people were not happy that there was a talk track that maybe the next move would be a cut. I think people are saying maybe now is not the time to raise rates, but the next move would be an increase in the rates. I wouldn't be shocked either way. I wouldn't be shocked if they hold rates again at the next meeting.

6:38I wouldn't be shocked if we see a 25 basis point increase. You know, you make a lot of really interesting points on the Treasury and Bessett's move and the yields and everything else. I saw this bullet this week, and I wanted to get your thoughts on this because I think this is – it just kind of shows the perplexing place that we sit right now. So the higher yields are magnifying the federal government's fiscal challenges, which you highlighted with deficits and everything else. Interest payments now consume nearly one-fifth of federal revenue, while publicly held debt is approximately 100 % of GDP.

7:10If you look at the Congressional Budget Office, they project federal interest rate costs will rise from 3.3 % of GDP this year to over 4.6 % by 2036, Stephen. Now, what's interesting is those projections are actually going to prove optimistic because they assumed a 4.1 percent 10-year Treasury yield, which is well below the current level, which this week touched almost 4.7 percent. And so according to the CBO, and this is a part that was just really interesting, just we talk about magnitudes of scale here. and the Congressional Budget Office said interest rates averaging just 10 basis points above its forecast would add$379 billion to federal interest expense.

7:51So for every 10 basis points, you're almost at$400 billion in additional interest expense. I mean, it's just hard for your mind to get wrapped around just how big, and to your point, just kind of this global massive need for capital and kind of deficits and spending and all the challenges that we're facing. I mean, when you start putting pen to paper on these things, it's just unfathomable. Yeah, I mean, the Social Security shortfall has been moved up now. I think we're talking about 2031. Initially, it was 2032, I think, somewhere in that neighborhood. So the reality is, by the time we get to the next major election cycle, 28, unless we have a candidate that stands a real chance of getting elected, that will actually tackle some of these tough austerity issues.

8:36we're in a terrible spot plowing into 2030. I mean, the amount of debt that's going to be rolling over at around that same point in time. Hopefully the AI build-out will be materially built out at that point in time, because heaven knows we need some good news and we need that compute to try and boost productivity. Because given our lack of appetite to tackle these difficult issues, one of the only remaining paths, and it's a dimly lit one. I don't see a lot of light shining through the end of this tunnel, but we have to work our way out of it. The productivity gains have to be so massive that somehow we can bail the water out of the sinking ship.

9:21But gosh, 2030 is looking like a really, really fast approaching, scary point in time for the U.S. economy. Yeah, I mean, I think that's the emphasis point is it's not that far away. You know, these numbers, 2030, 2036, they're not that far away. It'll be a blink of an eye and we'll be there. Something's going to have to break. We've been talking about this for a while now. Soft landing, no recession, no real economic pullback, government intervention, quantitative easing, you know, moves with Treasury, whether it be, you know, Pulte and them doing the stuff with Fannie and Freddie earlier in the year to now Besson doing this.

10:00So to your point, it's just a drop in the bucket. These are not fixing anything. These are just maybe sending a message. You know, maybe that's what they're trying to do. But if you go down the rabbit hole on some of this stuff, it can get pretty scary pretty quick. Oh, yeah, it can. I went down one of those rabbit holes this week. A buddy of mine sent me this academic paper called the AI layoff trap. And you want to talk about a dark rabbit hole to go down. said if AI displaces human workers faster than the economy can reabsorb them, it risks eroding the very consumer demand firms depend on. And so what this paper effectively shows is that if we can automate to the degree that, you know, the anthropic CEOs of the world have been talking about and that doomsday hits, the only way to kind of stabilize markets, get this, The only way in economic policy is through a tax.

10:56Rarely is a tax the optimal solution to so many of these economic problems. I mean, it is. There's a subset of them. But you basically have to have a Pigovian automation tax. tax and it's just like this really massive tax that basically says if you're going to lay off a worker because you're going to automate that work away we're going to tax the snot out of you because that's the only way we can kind of soften the blow it's like wow that's that's where we're going and i think we're going there sooner than what some people would think so whether treasury can actually put a ceiling on long-term yields is one question but the bigger theme here seems to be that there are a lot of borrowers competing for capital right now.

11:39We've been talking about that on every single show. And that makes this next story pretty interesting because we're also seeing firms look for different ways to unlock capital from assets they already own. Goldman Sachs announced it is acquiring LCN Capital Partners, a sale leaseback and net lease specialist in a deal worth up to$410 million. So Lonnie, walk us through this story. This is an interesting story, and I can't wait to hear Stephen's take on this, because this is a little unconventional relative to what we've seen over the last couple of months, both just in raw numbers and maybe strategic as well.

12:18So Goldman Sachs agreed to acquire LC and Capital Partners. They're a New York-based sale leaseback and net lease specialist for$260 million up front, and as much as$150 million of additional consideration, which are going to be tied to long-term performance targets. If you look at this, 80 % of the purchase price, or at least up to 80 % of the purchase price, will be paid in Goldman stock. And the transaction is going to be closed by the year end. It will add approximately$3 billion of assets to Goldman Sachs Asset Management. LCN is more than 40 employees, so pretty good-sized firm for this sale-leaseback type of structure, including co-founders Edward Lupuma and Brian York Colwell are going to join Goldman's real estate business.

13:04Now, just a little bit of background on LCN. They manage six portfolios, which consist of 375 properties across multiple asset classes, and get this, Stephen, across 20 countries. Its strategies include sale leaseback, as we noted, which also allow companies to sell their real estate, unlock capital, remain as tenants, as well as build-to-suit and triple net leases. So we've had a lot of anecdotal headlines around this kind of triple net, build-a-suit, sell-lease-back structure over the last couple of years as people are looking for a place to maybe place capital. But this competes directly with interest rates and some of the stuff that we just talked about.

13:43Goldman estimates the corporations own approximately$14 trillion of real estate across North America and Europe, only a fraction of which currently uses a net lease structure. So really, they're making a bet here on the structure. The firm sees considerable room for companies to monetize these properties as they seek alternative sources of capital. So I don't know if you have any thoughts on this, Stephen. I mean, I've always viewed kind of these sell leaseback or these net lease structures as somewhat risky in the sense that the buildings are either 100 % occupied or they're 0 % occupied. Yes, they pay rent and stuff like an annuity would, but that's assuming that they're still in business.

14:23And we've seen some challenges with some of these larger firms, CVS, Walgreens, et cetera, that kind of utilize this strategy. And now you're seeing it with Salad to Go and the bidding war that we're seeing with, you know, Dutch Bros and 7 Brew trying to buy some of their portfolio. So I don't disagree with the thesis here that there's a lot of value that could be unlocked, capital that could be unlocked. I am a little bit concerned, though, that this has somewhat of a binary outcome in my experience. It does, but it also has enough runway to continue expanding that I think it makes a lot of sense for Goldman Sachs to add a platform that has enough growth runway out there to make it look like a really attractive acquisition.

15:05I mean, this is a fairly modest transaction size for Goldman, but it could add some meaningful return to their bottom line. And ultimately, I mean, is all Goldman Sachs capital at risk? You know, I mean, so is it, it's not necessarily the binary outcome that's going to kill Goldman in the long run. They just have to bat an average of 300. You know, it's a very good, very strategic ad at this point in time, given what you're getting, the distribution platform, the global scale, and from what it sounds like a very, very strong, experienced, capable team. So I like this play, especially given how aggressive Goldman Sachs has been on trying to expand their asset management base.

15:48So it's a good one. It's not a massive needle mover, but it is right center of target on a fantastic inflation hedge asset. Yeah, I actually agree with you in this instance, because I think what they're doing is they've identified a company that has great experience, great execution skills, as evidenced by a portfolio of over 375 properties across 20 countries. But now with Goldman backing them and Goldman's platform and availability of capital and other things on their side, it really empowers LCN's team to do what they can at scale. And so I agree with you. And these are such small numbers. That's what I was saying on the lead-in is, you know,$260 million up front and$150 million additional.

16:34I mean, that's a rounding error for these folks. So I think it's a good strategic play for Goldman. We'll see how it plays out. Great news for LCN. And now they'll just have access to a much broader, more diversified set of resources, tools, and capital to go out and execute what they've proven to be a very strong play for them. Another story that caught our attention this week brings us back to AI, because somehow the numbers keep getting bigger every week. We saw in the Wall Street Journal that OpenAI has now signed a 20-year lease for a planned data center campus in Ohio, with NVIDIA helping backstop the value of the project.

17:14And the structure itself might be just as interesting as the size. So, Stephen, walk us through this one. So OpenAI signed a 20-year lease with SoftBank's SB Energy for a planned 10-gigawatt data center in Southern Ohio, potentially the largest AI infrastructure announced to date. And NVIDIA, as you mentioned, will initially backstop the completed value of the project's roughly 5-gigawatt first phase, helping SB Energy obtain less expensive debt financing by reducing lenders' residual value risk if OpenAI vacates. If OpenAI defaults or leaves, SB Energy must first attempt to release or sell the completed facilities.

17:54NVIDIA would then cover any value shortfall, potentially up to$105 billion. But its guarantee does not cover OpenAI's rent payments or facilities still under construction. In exchange, NVIDIA will serve as the exclusive chip supplier for the first phase, take an equity stake in SB Energy and potentially sell hundreds of billions of dollars of chips into the campus. NVIDIA also agreed to invest$1.5 billion directly in SB Energy. The revised arrangement is narrower than an earlier proposal under which NVIDIA could have guaranteed the entire 10 gigawatt project, potentially totaling$250 billion, and includes several mitigation steps intended to limit NVIDIA's exposure.

18:36SB Energy is reportedly preparing for an IPO that could raise$5 to$7 billion as soon as next month. The campus will be supported by a planned 9.2 gigawatt natural gas power plant owned by the U.S. government and financed by Japan under a recent trade agreement. So this is a really interesting structure where OpenAI supplies the long-term lease, NVIDIA helps support the asset's residual value, and sells chips. And those commitments allow the developer to raise enormous amounts of debt that would potentially be very difficult otherwise, certainly without that residual value guarantee. Steven, I don't want to be like Debbie Downer here, but when you start seeing NVIDIA, you know, effectively playing shadow support for all of these deals, you know, I remember I was talking a few months ago about someone that had made the infographic online where all roads lead back to NVIDIA, but it's just this circular capital flow.

19:33But when I hear SoftBank, I'm not like super pumped about that. Open AI, not super pumped about that. And NVIDIA, like this just screams to me like late 1980s savings and loans crisis where, you know, deals were getting appraised based off headlines, based on asking rents. It screams 2007 where, you know, issuance was being made based off underwritten financials and pro formas and rents that never were achieved. I mean, you know, I'm very bullish on the AI stuff, but some of this stuff, it just seems so circular in nature. I just don't know if I can get fully behind this stuff. I mean, like when every deal is backstopped by the same players and they're providing chips and they're doing this and it's how real is this?

20:17Like, can it actually withstand a major pullback here? Or is this all just headline fodder to get the deals done today with no real plan should some of these things not take hold like they think they will? I mean, it's essentially a go for broke, all or nothing play. NVIDIA is doing the smart thing, right? They're basically boxing out any competitors here, saying nobody else is going to be able to sell chips into this campus. This is a massive facility and NVIDIA needs to be the exclusive provider. So they're essentially ensuring that as long as we stay center of the fairway, more or less, the amount of revenue they're going to get off this project will be massive.

21:00And yeah, if they lose everything on the project, I mean, they kind of break even, right? They've sold the chips, they have the guarantee. And so it ends up being sort of a net wash. But on the more positive scenario end of things, they still come out probably a couple hundred billion ahead. on on net but it's a massive guarantee they're taking on here look it's what do you learn when you're growing up it's like you don't co-sign on other people's stuff right i mean it's just a recipe for bad bad news but nvidia is being the backstop not just on this deal if this was a one off deal then i would say maybe this is worth the risk but it feels like on all of these deals they have some sort of backstop guarantee some sort of whatever and yeah from the competitive perspective I think your point you make is valid.

21:43They're kind of boxing others out. But it just gets interesting for me. Like how many of these deals can they effectively do if they go like, what is the critical mass on these things going bad before it breaks? If it's just this one in isolation, probably not a huge risk. But if they have five or six of these where they're playing this role and they all don't work or they all slow down or open AI, you know, goes default, whatever, how does that work? Like that's the part I just, is anyone even paying attention? to that? That's the part I'm not sure of. Right. Because it looks very, very, very much like an AIG structure.

22:20What brought down the entire market in 07, 08 was that inner leakage through AIG selling a massive amount of default version, correct default swaps. They were the central counterparty. When you take them out of the chain, entire system falls apart. And the same thing could happen here but really is nvidia does go belly up i mean the rest of us are probably eating dog food in that economy yeah that's part of their strategy steven is it's a too big to fail and i think there there's some comfort that the government wouldn't allow it to happen and i don't know how i feel about that either now what i want to know is on the legal side i really want to review one of these agreements and i want to see how quick has nvidia gotten on the legal structuring of these, does it look sort of like an insurance policy that's designed to not pay out?

23:16Like, have they crafted it that well? Or is it truly as generous as it sort of appears at face value? Hopefully these deals work and we don't have to get into those details. But, you know, the reality is, I think a small number or a certain number of these deals are going to have some headwinds. And we'll see at some point how they play out. I mean, it's, I love that people are willing to take the risk. I mean, that's what innovation requires. And that's how you make progress in the world. You have to have people willing to bet it all and kind of, you know, take the risk. So I'm not negative on that part of it.

23:50But I just, you know, the logical brain of me says some of these things, just everything has to go right. And in reality, rarely does everything actually go right. And we have one more AI story. This comes from the Wall Street Journal. They looked at nine major tech companies and found roughly$3 trillion in off-balance sheet commitments, largely tied to AI infrastructure and far beyond what shows up in their reported CapEx numbers. What do you guys make of this? Well, speaking of too big to fail, I love this announcement because it gets to the issue in the show, but puts some firm numbers to it.

24:31The balance sheets represent what's in existence today to some degree, but isn't capturing the forward commitments and some of the financing structures that we'll talk about on the show. So the nine major technology companies have approximately$3 trillion of off-balance sheet commitments, primarily tied to AI infrastructure. That compares with roughly$600 billion of reported capital expenditures over the past year, and about three times their combined recognized leases and long-term debt. That consists of approximately 1.2 trillion of leases that have been signed but have not started and 1.9 trillion of contractual purchase commitments for chips, equipment, power, and other infrastructure.

25:14These obligations generally remain off balance sheet until rent begins or the contracted product or service is delivered. Alphabet, Amazon, Meta, and Microsoft collectively account for roughly 4 of these commitments, $904 billion of future leases and$1.52 trillion of purchase agreements. Alphabet alone reported approximately$902 billion, followed by Meta at$696 billion, Microsoft at$558 billion, and Amazon at$267 billion. The obligations are also expanding much faster than traditional capital spending. Alphabet's off-balance sheet commitments increased by more than 800 % from a year earlier. Metas rose by 750%, and Oracles grew by more than 500%.

26:07Alphabet's purchase commitments jumped from$332 billion to$811 billion in just three months, months, including technical infrastructure, inventory, and data center agreements, extending as far as 2054. Meanwhile, Meta has 347 billion of leases that have not commenced, including obligations associated with its Hyperion Data Center project in Louisiana. So the optimistic case is that AI adoption produces sufficient revenue to cover these commitments. The risk is that locking themselves into largely non-cancelable leases and purchase agreements, just as heavy capital spending is pressuring free cash flow.

26:50For investors and lenders, reported debt increasingly understates big techs, effective leverage, and fixed financial obligations. Yeah, I love being able to get some of this context, Stephen. I think this is a nice segment relative to what we just talked about before. And, you know, look, these tech companies are innovative. They have piles of cash. Obviously, they're leveraging up on some of these deals here. I'm not betting against the seven, I think we call nine major tech companies in the article or the big seven that we generally talk about. But it is good to kind of get a little bit of quantification of kind of where these things sit and how much the exposure is, even if it's not, you know, exact.

27:34It's an interesting thought exercise. And listen, you were talking about academic papers earlier today on automation or whatever. I think we're on the precipice of having a lot of people just do some deep dive analysis on the risks of just this increasingly hyper leveraged type of AI infrastructure, spin, development, R &D, all of this stuff. I'm hopeful that we start seeing some real results from the AI, which we are all banking on, because these numbers are just pretty big. Quick break. There are a lot of AI tools entering the commercial real estate market, and Lev is one of the platforms taking a more practical approach to putting AI to work for CRE teams.

28:19It's not just financing software anymore. Lev gives you ready-to-use AI apps and agents for research, comps, deal materials, outreach, documents, pipeline, and team workflows. Go to lev.com, that's L-E-V dot com, create an account, get$2 ,000 of free credits, and tell them Jake sent you for another$500 in credits. All right, so let's bring this back to some commercial real estate data and dig into TREPS data specifically. We've talked a lot about office distress through the lens of delinquency, special servicing, and other metrics. But this week, we have an analysis that looks at another group entirely.

29:04Loans that are still making their payments, but whose properties are no longer generating enough cash flow to actually cover the debt. And the surprising part is that vacancy doesn't explain all of it. So, Stephen, kick us off here and let's dig through the data. What did you find when you looked at these performing office loans with DSCRs below 1.0x? Right now, we have$97.2 billion of performing securitized loans on conventional urban and suburban office buildings. These are seasoned loans with recent financial reporting that remain current on their debt service payments and have not been marked as delinquent.

29:46But performing doesn't necessarily mean financially healthy. Within that population,$12.1 billion across 162 loans are roughly$1 out of every eight of these performing securitized urban and suburban office loans. almost one in every$8 has a debt service coverage less than 1.0 times, which means the property's annual cash flow cannot cover its annual debt service payments. That's a huge problem, right? That basically begs the question, do we have about 12.1 billion of office loans that are going to become delinquent? So the first blog that we put out here on this population of office loans, with the debt service coverage less than 1.0, performing office loans with a debt service coverage ratio less than 1.0, is looking at what's actually causing the shortfall.

30:41Now, the obvious answer that would pop into pretty much most of our heads would be vacancy. And that would seem to explain at least a good portion of this. Approximately$6.9 billion, or 57 % of that 12.1 billion, has an occupancy less than 80%. The median occupancy rate for that group is only 62%. And its median DSCR is 0.7 times coverage. But what's surprising here is that means that the remainder of this population has an occupancy above 80%. 5.2 billion, or nearly 43 % of that 12.1 billion, has an occupancy rate of 80 % or higher. And the median occupancy within that group is 93%. And yet its median DSCR is still only 0.78 times coverage.

31:33So in other words, the tenants are largely in place, but the cash flow is not reaching the bottom line. So what we did is we did a little bit of fuzzy text analysis here in servicer comments to see if free rent is explaining at least part of this population. And sure enough, we were able to identify 1.46 billion or roughly 28 % of that well-occupied group has some sort of free rent issue in place. But here's the kicker. One massive loan accounts for a huge chunk of that. Just over$1 billion comes from 280 Park Avenue alone. So free rent is not a complete explanation of that broader question of what's causing the low debt service coverage for these well-occupied buildings.

32:22Now, we dug a little bit further and found that, well, Well, it's likely a combination of, first off, a chunk of these are floating rate loans. So higher debt service costs are accounting for part of that pain. And then expense creep is accounting for a large chunk of the other. The operating expense ratio for that well-occupied group is running about 65%. I mean, that is a massively high OPEX ratio for the well-occupied group. So expense creep has been extremely, extremely painful. Now, to just show how much the free rent can drain on cash flow, if we look at the$200 million loan on 59 Maiden Lane in Manhattan, they reported full occupancy, so 100 % occupied, but a DSCR of only 0.33 times at year-end 2025, because five months of free rent ran through mid-December.

33:13Once that concession burned off, first quarter 2026, DSCR jumped to 2.65 times. So it is taking a lot of concessions to fill these buildings up, Once those concessions burn off, you have a lot more light at the end of the tunnel, but it takes some financial pain to get to the end of that tunnel. So this brings us to the second blog we put out, which deals with when these borrowers will have to confront the shortfall. In other words, what does the maturity schedule look like? So Lonnie, do you want to walk us through what that maturity schedule looks like? Yeah, happy to do that, Steve. And I did want to add, you know, just a great reminder in your segment there on how physical occupancy only tells one part of the story.

33:51You have to look at those concessions in all property sectors. And you think of concessions generally being most prevalent in multifamily. But in the office sector, to your point, to get occupancies up, they've had to give away a lot of free rent. So it's amazing to see. That's a great data point to say when rents or the concessions are in place, you know, 0.33x, you know, as soon as the concessions jump off, it jumps to 2.65x on the DSCR. I mean, that's a real tangible example of how impactful those can be. On the maturity side here, of the$12.1 billion,$5.66 billion still have extension options available, which is an important note.

34:27$6.44 billion have no extensions remaining. The group with extensions consists of only 16 loans. They have an average balance in place of approximately$354 million. The 148 loans without extensions average just$44 million, so much smaller balance per property. So much of the extension exposure is concentrated in a relatively small number of very large loans. If you assume every available extension is exercised,$11.1 billion, or 91.8 % of the entire sub-breakeven balance, reaches a hard maturity by the end of 2029. A hard maturity, as you've defined multiple times for us on the show, Stephen, is the point at which no contractual extension options remain, which requires the loan to be repaid, refinanced, or renegotiated.

35:10The largest wave arrives in 2028 when$4.73 billion reaches hard maturity, represents about 39 % of the total$12.1 billion population. Much of that concentration comes from five large loans, exhausting their extension options all in that same year, including three with balances of at least$950 million. Extensions, therefore, provide time. But as we've talked about multiple times on the show, Stephen, extensions don't eliminate the problem. The$1.2 billion loan at San Francisco's 555 California Street campus illustrates that point. If you look at it, DSCR fell to 0.47x in the first quarter of 2026.

35:50Although the loan still had extension options, it was renegotiated in May and remains current, showing that borrowers and lenders may need to act before the contractual deadline arrives. If you look at the broader takeaways here, delinquency rates alone provide an incomplete view of office credit. these 162 loans are all still performing today, but the properties are not generating enough cash to support their debt. I mean, that's a pretty interesting construct where they're not delinquent, but they're not producing enough cash flow to even cover the debt service. Some may improve, as you highlighted, Stephen, when the free rents expire.

36:22Others may face more structural pressure from vacancy or these operating expenses that have creeped up over time or the floating interest rate expenses that you highlighted as well. If you look all the way out by 2029, nearly 92 % of the imbalance will have to be resolved. Payment status tells us which loans have already encountered trouble. If you look at DSCR, maturity schedules, they help us identify which loans may be next. This is a great exercise for us to kind of go through. And you've done a lot of work on the hard maturity date stuff, Stephen. We've also done a lot of work that highlights there's a lot of loans that are past their maturity date that have not been able to refinance, have not been able to reach some sort of resolution and have not been able to sell.

37:04So I think the term hard maturity is still somewhat of a misnomer, right? In the sense that the lender is not effectively taking these buildings back the day after the hard maturity, they're working through these problems even beyond that date. So more to come on this. We'll be tracking these properties, obviously. And I know, you know, you had done a little bit of a breakdown of just driving the coverage shortfalls. And we call it hard maturity on a soft footing. Yeah, I think that's a fair representation. So if you're interested in any of the data that we just ran through, reach out to us at podcast.trip.com.

37:42We'd be happy to share more details or answer any questions or show you how you can access all of this data and insights within our platforms or via our MCP server or any other way that you'd like to get access to our tools and data. Turning here to our deals and data segment and some property type stories, we wanted to talk about a few sales today. There was a headline in BizNow that Rexford is set to sell a$1.2 billion industrial portfolio as part of its plan to increase dispositions. Yes. Southern California-focused Rexford Industrial Realty agreed to sell a$1.2 billion industrial portfolio to an EQT real estate affiliate, with closing expected by the end of the third quarter.

38:26The transaction will bring Rexford's 2026 dispositions to approximately$1.5 billion, the bottom of its revised$1.5 billion to$2 billion target, and a sharp increase from its initial$400 to$500 million guidance. Proceeds will help address$1 billion of debt maturing in 2027, equivalent to roughly 30 % of Rexford's total debt. The REIT also plans to fund share repurchases and development investments. The portfolio carries an estimated 2027 net operating income cash flow yield of 5.5%, incorporating expected tenant move-outs and declines as above-market leases reset to lower rents. Following a portfolio review, Rexford identified approximately 8 million square feet or 2 billion of assets that no longer fits its long-term strategy.

39:16That represents about 16 % of its 50 million square foot portfolio. The target properties generally offer limited long-term growth potential or face risks from above market rents, short lease terms, and elevated competing supply. Accelerating the sales required Rexford to recognize a non-cash impairment, contributing to a$506.9 million second quarter net loss attributable to common shareholders. So this sale illustrates how even a major industrial REIT is prioritizing liquidity, debt reduction, and portfolio quality over maintaining scale. It also highlights the pressure that rent roll-downs and near-term maturities can create despite industrials' comparatively strong long-term fundamentals.

40:01And let's talk about some other sales we saw. This is in the lodging segment. California hotel sales reached$1.63 billion during the first half of 2026, which is up 17.2 % year over year. But the increase largely reflects distressed and lender-driven transactions rather than broad market strength. Yes. Lender actions preceded 36 of the state's 128 hotel sales. The transactions represented 28 % of deal count and 37 % of dollar volume. The refinancing wall is expected to generate additional distress. 30 % of U.S. hotel loans mature during 2026, and many were originated when rates were around 4%.

40:47Refinancing today frequently requires owners to contribute substantial new equity. We call that a cash-in refi. Expensive brand-mandated renovations, rising construction costs, and higher labor expenses are adding to the pressure. The West and Long Beach, for example, reportedly required a$25 million property improvement plan before selling for$42 million, less than half of its 2017 price. Other distressed transactions included the addition West Hollywood transferred through a deed-in-lieu with$211 million of debt, and Napa's Stanley Ranch acquired by Blackstone following a$220 million default.

41:26Public hotel REITs have largely shifted from buyers to sellers as they've reduced leverage. Owner-operators, private equity firms, multi-individuals, and hotel-to-housing investors are filling the gap. Activity varied sharply by region. Southern California sales volume jumped 57 % to just under$1 billion, while Northern California volume declined 15%, almost 16%, to roughly$650 million. Investors increasingly view San Francisco and Silicon Valley as recovery opportunities, supported by AI investments, employment growth, and improving business travel. Los Angeles continues to lag because of its slower economic recovery and weakness in film and television production.

42:07California's rising transaction volume is functioning as a mechanism for price discovery and balance sheet repair. Maturities and capital needs are forcing sales at recent valuations, while well-capitalized buyers selectively acquire assets where they see credible demand recovery and limited new supply. You're always going to have that high barrier to entry in California, maybe not as much in lodging as you do in multifamily, but I do think there's going to be some really good opportunities here. I mean, it comes back to, especially in LA, where the market is a little bit slower to recover, you have to be well-capitalized to be able to buy these assets, even at a reduced basis, and manage them for the next however long, several months to year, at lower than market occupancies.

42:48But I think long-term, there's some real value here, just because LA is not going anywhere. Obviously, San Francisco has already turned the corner. So it's good to see. I'd like to see these volumes continue to increase. Obviously, you don't want them driven by distressed sales. But as we've noted, Stephen, that's the one component of the market that we've been missing. And so it's good to actually see that the logjam break a little and some of these transactions start to take shape in the market because it does provide price discovery. And it does mean that you're actually exercising that part of the cycle, which leads to broader recovery.

43:20So I take this story, even though catalyst for these sales may be distressed, the outcome, I think, is positive for all involved. So before we close, I have some programming notes. If you're listening to this episode, that means you just missed our TREP MCP, Connecting AI to Your Models, Your Processes, Your Workflows webinar that we hosted. We had so many of our listeners reach out, and we hope a lot of you got to join that. But if you missed it and you're curious about what we showed, reach out to us at podcast.trep.com. And we'd be happy to book a specific call for you to actually walk you through your use case.

43:56And if you were on that webinar and you feel like you want to see this in action and in practice for your team, reach out to us. We'd be happy to get on the phone with you. We'd love to do a quick run through of what your day-to-day looks like and how our tools and data could really improve and expedite some of your processes. But even if you missed that, you still have a chance to sign up for our upcoming MarketPulse webinar. This next event is taking place on Thursday, August 27th at 2 p.m. Eastern. And this month, we're covering line item financials. We'll show you how property revenues operating expenses are shifting across CRE, which costs are rising fastest and where the greatest pressure on net operating income is emerging.

44:37We'll walk through cash out refinancing trends and explore how the rise in noteworthy equity distributions from newly securitized loans in 2026 and how today's cash out activity compares with years prior. And we'll also be giving a back to school special. So if you're interested in an update on the student housing market, we'll walk through the latest performance trends and lending activity. So email us podcast.trep.com or check your email if you get our TREP emails or visit our LinkedIn page and you can find the link to sign up for that webinar. Turning to shout outs, Josh M was interested in joining our MCP webinar.

45:14Chris S thanked us for the great info on a recent analysis we put out that said that talked about what's behind the July multifamily CMBS delinquency uptick. Brandon D was interested in our 101 piece on back leverage. He said, also love the podcast, listen to it every time I go to work. We had a lot of people interested in that back leverage section that we did a few podcasts ago and we do have it in written form if anyone's interested. Tritee is a listener and mentioned that to one of our colleagues on a recent demo so thank you for listening to the show. Jason R was also interested in the MCP webinar and was interested in seeing how TREP data can connect with different AI tools.

45:55Michael D had a similar request and thanked us for the show. Kevin M said great show this week and wanted to see our latest CMBS special servicing rates. Sarah L. said she's very focused on multifamily performance and thanked us for a great show and wanted to see what other resources we had. Daniel H. is a longtime listener, but said it's his first time reaching out. He really appreciates our digging through the data sections and told us to keep up the great work. And Andrew C. was interested in the latest delinquency rates that we had and would like to take a closer look. So as always, thank you to our listeners for reaching out.

46:29And one more shout out here before we go. We wanted to shout out to the team at California MBA. They are hosting the 2026 Western States Cref Conference, September 9th through 11th at the ARIA in Las Vegas. And your very own right here, this podcast crew, will actually be the keynote luncheon. We'll be doing a live recording of the Trapwire podcast at this conference. We are so excited. We did this a few years ago. And if you're interested, you can actually find that show online. but we're really grateful to the California MBA team for putting on such an interesting and engaging conference that really gets to all the root of what's going on in commercial real estate, lending, advocacy, policy issues.

47:12And we're really excited to meet with attendees there and do another live show. So that will be an exclusive for everyone at that conference, but we will be recording it and releasing it so our listeners can hear it as well. If you're in Vegas or you're going to be at that conference, send us a note. We have some time there. We'd love to meet with you and connect and hear what you're working on. And with that, we'll close. Thanks to our producer, Mariana Sobrana. Join us next week as we look at what's happened during the week and how it may be impacting you. If you have a question or just a comment, send an email to podcast at trep.com and subscribe to the Trep Wire podcast with your favorite provider.

47:49Thank you for listening and stay well. All right.

47:59We'll be right back.