Blackstone Mortgage Trust (BXMT)

Business Overview

Blackstone Mortgage Trust does not own buildings. It lends against them, writing senior floating-rate mortgages across the US and Europe, and it sits above the owner in the queue when collateral fails. There is no agency MBS portfolio and no servicing platform, so nothing here turns on prepayment speeds or hedging. The 2025 originations went out at an average 66% loan-to-value, which is 34 cents of somebody else's equity underneath every dollar lent, measured against the appraisal of the day. That cushion is the business. It holds only while the appraisal does.

Market capitalisation was roughly $3.10 billion as of 10 June 2026, on 168.6 million shares outstanding at an $18.41 close. Book value per share was $20.75 at 31 December 2025, already net of $1.76 per share of CECL reserves, the lender's own estimate of losses it expects on the loans it holds. The stock traded at a discount to that book, below the 0.90× line that sector convention treats as a deep discount, where the market is pricing further book erosion rather than a wobble.

Two earnings figures describe the same year. FY2025 distributable EPS, the cash-basis measure the dividend is declared against, was $1.86 per share before credit charge-offs and $(1.43) after them. The declared dividend was $1.88 ($0.47 a quarter). On the first line the payout was almost exactly covered; on the second it was paid out of book value. Everything about how the stock is priced follows from which of those two lines an investor believes is the run rate.

How CRE Lending Economics Work

Loans receivable were $17.8 billion net of reserves at year-end, spread across 131 loans. They are floating-rate, which lifted net interest income while base rates were high. A floating coupon does nothing for the value of the collateral, though, and that is the exposure that matters here: the loan reprices, the building does not.

Office is where BXMT diverges from the diversified lenders, and it is also where the numbers most often get compared wrongly. Blackstone Mortgage reports office at 27% of net loan exposure, split 20% US and 7% non-US. Starwood Property reports US office at 8% of undepreciated assets, but that denominator is the whole company: commercial lending, infrastructure lending, buildings it owns outright, servicing. Dividing the same dollars of office by a far bigger number produces a far smaller percentage, so the two figures cannot be ranked against each other. What does compare is the reserve each carries against book: $1.76 a share inside a $20.75 book at Blackstone Mortgage, about 8%, against $1.84 inside a $19.25 undepreciated book at Starwood, about 10%. Two points apart, on credit disclosure that looks nothing alike. The CRE debt guide works through why.

The impaired loan balance was $0.1 billion at 31 December 2025, down 96% from a 2024 peak of $2.3 billion, and the company reported 99% of the loan portfolio performing. Neither figure says the credit problem went away. Charging a loan off removes it from the impaired balance, and BXMT resolved roughly $0.6 billion of impaired loans in the fourth quarter alone. The $3.29 a share between distributable EPS before and after charge-offs is where those loans went. A small impaired balance is what a book looks like after the losses have run through it, not instead of them. The performing percentage carries its own caveat: extending or modifying a loan keeps it performing on paper, so 99% is only reassuring once you know how much of the book got there by being given more time.

Distributable EPS and the Dividend

Distributable EPS is the non-GAAP measure Blackstone Mortgage declares its dividend against. It strips out depreciation and unrealised marks but includes losses once they are realised, which is why it splits into two lines in a year like 2025. Before charge-offs it was $1.86 a share, covering the $1.88 dividend 0.99× ($1.86 ÷ $1.88), thin but positive. After charge-offs it was $(1.43), and coverage arithmetic stops meaning anything: the payout came from capital.

The company held the $0.47 quarterly dividend into 2026, which says the board expects pre-charge-off earnings power to return, or is willing to bridge the gap from the balance sheet for a while. But this rate is already the rebased one. The dividend was cut from $0.62 to $0.47 in the third quarter of 2024, a 24% reduction, as the office reserves first went up. What is being defended is the second level, not the original.

Book value of $20.75 already carries $1.76 a share of CECL reserves, but a reserve is a management estimate rather than a realised loss, and it moves in both directions. Total CECL fell 60% during 2025 to $296 million, including $33 million released back into earnings on loans that were resolved above their carrying value. Further collateral marks would run the other way, hitting distributable EPS and book in the same period. The book value guide uses BXMT to show how CRE credit events compress both the earnings and the multiple paid for them.

Valuation Framework

A discount to book at a lender is neither a bargain nor a warning by itself. It is the market's estimate of how much further book value has to fall, so the useful thing is to weigh it against what management has already taken. The discount on 10 June 2026 was worth roughly another reserve on top of the $1.76 a share already inside book. For that to be right you need a second charge-off year on the scale of 2025, or office collateral marking below what the reserves assume. For it to be wrong you need distributable earnings back near $1.86 with the reserve released rather than consumed. The discount settles neither question; it only tells you which way the market is leaning.

Everything else hinges on which distributable EPS line you treat as run-rate. On $1.86 before charge-offs, the dividend is nearly covered. On $(1.43) after charge-offs, there is no multiple to compute, and the dividend is coming out of capital.

Office cap rates and transaction evidence drive the forward marks, and the cap rates guide gives the property-type context for stress-testing a book with 27% of its net loan exposure in office. Starwood is the diversified comparison, with distributable earnings covering its dividend 0.88×. BXMT is the concentrated-office case inside the same sector label.

What to Watch in the Financials

Distributable EPS bridge. Each quarter separates recurring net interest income from charge-offs and provisions. The $1.86 becomes a credible run-rate only if new charge-offs fade.

Dividend policy. $1.88 a share annualised against negative after-charge-off earnings is the live tension. A second cut would say the charge-off year was structural rather than a clear-out; holding the rate while pre-charge-off earnings recover would support the bridge.

Where the loan book is going. Office was 27% of net loan exposure at year-end, against 50% in multifamily and industrial combined. Repayments in 2025 ran $6.1 billion, of which $2.3 billion was office, and Q4 originations were entirely multifamily and industrial. Watch whether that rotation continues, or whether office loans stop repaying and stay on the book by default.

Reserves and how they move. The $20.75 book already embeds $1.76 a share of reserves. Watch whether releases come from resolutions above carrying value, which is a genuine recovery, or from reclassification, which is not. New provisions hit book and distributable EPS in the same period.

Key Risks

Uncovered dividend after charge-offs. Distributable EPS of $(1.43) did not support the $1.88 paid. A payout continued without earnings behind it draws on book value or on new capital, and the rate has already been cut once.

Office concentration. Office is 27% of net loan exposure, and cap rate expansion or vacancy stress lands on that slice first. The exposure is not office as a category. US office vacancy has been improving, but the improvement is concentrated in prime buildings, and the lender's problem is the older stock behind them: an owner who cannot fund the fit-out a tenant now demands has a building nobody will refinance, and the loan against it has nowhere to go.

Charge-off recurrence. The $1.86 before charge-offs shows the spread business still earns. The question is whether 2025's losses were a clear-out or a run rate.

Ageing appraisals. The 66% loan-to-value on 2025 originations is struck against the appraisal of the day and does not update itself. Blackstone Mortgage carries the office property it has repossessed at roughly half what those buildings were appraised at when the loans were written, which is the scale of re-marking a stale valuation can hide.