Castellan Real Estate: The Funding-Line and Encumbrance Map Behind the Book

What Exactly Is Being Isolated Here

The main article argued that Castellan screens as conservative at first glance: a $471.117 million loan book, a 49.7% average LTV, and no reported principal losses. This follow-up isolates the next question, and it is a different one: through which channels is that book funded, how much of it is already ring-fenced and pledged into those channels, and how much value is still genuinely free at the top of the structure.

The number that stands out in the presentation is $532 million of funding lines, of which only $189.2 million was drawn and $342.8 million remained unused. That sounds like abundant flexibility. But the same evidence set shows something less comfortable: $428.442 million of the relevant asset base was already pledged, or 91% of those assets, while only $42.675 million remained unencumbered. This is not a story about a shortage of collateral. It is a story about access to collateral.

There is also an important gap between two layers that a reader can easily blur together. At the underlying real-estate layer, the presentation points to roughly $961 million of pledged real-estate value against only about $312 million of net debt at the fund level, implying a roughly $649 million cushion. That cushion is real. But at Castellan's own balance-sheet layer, the encumbrance table shows that most of the loan book is already sitting inside pledged funding structures. For bondholders, the practical question is therefore not only how thick the collateral cushion is at the property level, but how much of the book remains unpledged and how much of the cushion already sits behind other lenders.

That chart immediately shows what is hiding behind the $342.8 million unused-capacity headline. This is not an evenly distributed reserve. Most of the apparent flexibility sits in one line.

The Funding-Line Map: Diversification Exists, but Spare Capacity Is Concentrated

At the name level, Castellan really does have five separate funding relationships: Axos, Western Alliance, Lancewood, Northeast, and Churchill. This is not a one-lender structure. The outstanding balances are also not dominated by a single line. Axos stood at $59.6 million, Western at $52.3 million, Lancewood at $33.8 million, Northeast at $25.8 million, and Churchill at only $17.7 million.

But once the question shifts from diversification to spare capacity, the picture changes. Of the $342.8 million unused, $232.3 million, about 67.8%, sits in Churchill. Northeast, by contrast, was already 80.6% utilized. Axos and Western were both around 58% to 60% utilized, and Lancewood was around 56%. So the true reserve is not five half-empty lines. It is essentially one large repo line that had barely been ramped by year-end.

There is another non-obvious point here. Section 10.5, which aggregates only bank credit, shows bank lines of $222 million and utilization of $137.8 million. That means only about $84.2 million of the unused capacity sits in the bank lines. The remaining unused capacity, about $258.5 million based on the detailed facility figures, comes from the two non-bank facilities, Lancewood and Churchill. Anyone reading the $342.8 million headline as broadly available bank capacity is reading the map too flatly.

This chart produces the first core insight of the continuation: the space to fund growth or replace loans is not truly spread across five equal relationships. It depends heavily on Churchill's willingness to scale a relatively new repo line.

What Non-Recourse Means Here, and What It Does Not Solve

The presentation states the point cleanly: each facility is backed by a designated loan pool and is non-recourse to the company. That is true, but it matters to unpack what it means. The business is funded through dedicated subsidiaries. Each relevant subsidiary pledges its bridge loans and related assets to its own funding counterparty. In that structure, trouble in one line is not supposed to pull the public parent automatically into the same event. That is a real structural protection.

But non-recourse does not mean the assets are free. Quite the opposite. Section 10.1 explains that in an event of default under a funding agreement, the lender may redirect borrower payments to itself. In some cases, the subsidiary can also be restricted in its use of cash received from the loans and may be required to use that cash to repay the funding line. If the end borrower defaults, the lender may have the right to enforce remedies directly against the bridge loan and even trigger cash-trap mechanisms inside the funded pool.

That is the key distinction between risk that does not automatically travel back to the parent and flexibility that can actually move upward. The structure does isolate each line. But it also isolates each line inside its own pool of loans. So the $342.8 million of unused capacity is not the same thing as parent-level free liquidity. To access it, Castellan still needs to originate loans that the relevant lender will approve, inside the relevant entity, at the leverage terms that the specific agreement allows.

There is a second layer inside the documents themselves. Note 8 describes a line-by-line approval mechanism, with advance structures that, where disclosed, sit around 65% to 80% of principal and 50% to 60% of collateral value. The facility summary in section 10.2 also shows that the support structure is not uniform. Some lines carry limited guarantees or bad-boy carveouts tied to sponsor-related entities, and some include full guarantees from related entities. So even if the presentation summarizes the setup as non-recourse to the company, the funding lines themselves are not always clean asset-only pipes. In some cases there is also a sponsor-support layer behind the borrowing entity.

The Encumbrance Map: A Big Cushion Is Not the Same as Free Assets

This is the heart of the article. The presentation quite rightly highlights the cushion implied by an average 49% LTV, roughly $961 million of pledged real-estate value, and only about $312 million of net debt at the fund level. That matters, because it says the asset layer underneath the loans is thick.

But a Castellan bondholder does not sit directly on that real-estate layer. The bondholder sits above subsidiaries that fund those loans through pledged facilities. And section 10.8 sharpens that point without rounding away the detail: at year-end 2025, there were $428.442 million of encumbered assets, representing 91% of the relevant asset base, while only $42.675 million was unencumbered, or 9%. Against that, secured liabilities on company assets stood at $269.257 million.

The key calculation here is not only how much debt exists, but where it sits. Between $428.442 million of encumbered assets and $269.257 million of secured liabilities lies about $159.2 million of value inside the pledged pools. That is a real residual layer, but it is not the same thing as free assets. For that value to move upward, it first has to survive every senior funding layer inside every ring-fenced entity.

So the encumbrance map has to be read on two floors:

• At the property layer: there is a large cushion against the collateral value supporting the borrowers.
• At Castellan's own bond-issuer layer: only about $42.7 million of the book remains directly unencumbered, while most of the book is already financed through senior facilities.

The presentation itself bridges those two readings when it points to roughly $43 million of unencumbered loans. In other words, the "free" slice highlighted in the presentation is almost exactly the same slice that the annual encumbrance table marks as unencumbered assets. That is not a trivial number. But it is also not a number that lets a bondholder ignore how tightly the book has already been mapped into secured financing structures.

Why the Runoff Mechanism Matters More Than the Headline About Refinancing

It is easy to see the encumbrance map and jump straight to a refinancing-dependence conclusion. That is not fully right. Section 10.1 and the presentation both emphasize that if a facility is not renewed, the repayment mechanism is not a classic bullet wall that has to be taken out on a single day. The pledged loans themselves repay the line as they repay in practice, for a year after the facility term ends. Churchill adds a different protection: it has no fixed maturity, but the lender must provide one year's prior notice if it wants to terminate the arrangement.

That matters because it changes the type of risk. Castellan is not built around one warehouse that can simply shut tomorrow morning. The system is built so that the funding lines amortize through the assets they fund. So the true refinancing question is not only whether a renewal occurs. It is also whether the underlying loans keep repaying, selling, or rolling in a way that allows that runoff to happen in an orderly fashion.

In that sense, Castellan's structure does two opposite things at once:

• It reduces the risk of a single hard refinancing wall at the parent level, because facility repayment is linked to loan repayment.
• It increases dependence on the operating quality of each funded pool, because cash that gets stuck or trapped in that pool is not freely available somewhere else.

As long as the book continues to behave like a short-duration bridge-loan book with low LTVs, that structure can work well. If the extension layer, delinquency layer, or loan-sale layer starts widening, the same structure can shift very quickly from a mitigant into a constraint on flexibility.

Bottom Line

The message of this continuation is simple: Castellan has a better funding map than the headline "small bond-only issuer" might imply, but a tighter free-asset picture than the headline "$649 million of cushion" might imply. Both are true at the same time.

On the one hand, five funding lines, $532 million of total capacity, only $189.2 million drawn, runoff mechanics instead of straight bullets, and a large Churchill line that was barely used all paint a funding picture that is less stressed than a superficial read might suggest. On the other hand, 91% of the relevant asset base is already pledged, the unused capacity is concentrated mainly in one line, and most of the value sits inside ring-fenced entities where the funding counterparties can control cash in stress.

For a bondholder, the practical conclusion is that it is not enough to ask whether a cushion exists. The real question is where that cushion sits. In Castellan's year-end 2025 structure, a meaningful part of the cushion sits first below the facility lenders and only then above them.

Three monitoring points follow naturally:

• whether Churchill actually becomes a line that funds a larger share of the book in practice, rather than remaining mainly a theoretical ceiling that enlarges the sense of flexibility;
• whether the pool of unencumbered loans stays around the $43 million area or erodes as more assets get painted into facilities;
• whether the first real cash-trap event, tightening of terms, or decline in lender approval capacity appears in one of the main lines.

If those three points stay clean, Castellan's funding map can read as a genuine strength. If they do not, the market will start to see that the central question is not only the quality of the underlying real-estate collateral, but the quality of the path between that collateral and the bondholder.