BLACKSTONE MORTGAGE TRUST, INC. (BXMT) stock intrinsic value fair value estimate & owner earnings analysis

Blackstone Mortgage Trust is a commercial real estate lender managed by a subsidiary of Blackstone. It makes large, mostly floating-rate senior loans on office, multifamily, industrial, hotel and other properties in the United States, the United Kingdom, Ireland, Australia, Spain, Sweden, Canada and elsewhere in Europe, funds them mostly with borrowed money, and pays out most of what it earns. At June 30, 2026 it held 133 loans with $17.4 billion of principal plus about $1.3 billion of foreclosed property, against $16.1 billion of liabilities and $3.26 billion of shareholders' equity, or about $19.35 per share. This company was Capital Trust, Inc. until Blackstone took over its management in December 2012. For a lender the main long-run cost is loan losses, and this analysis is built around them. How we measure earnings. We start from cash generated by operations, remove short-term swings in receivables and payables, and deduct stock-based pay. We add back the cash origination fees borrowers pay up front, which the company records outside operating cash flow; they averaged $63 million a year over 2016-2025 and we use about $60 million (about $0.36 per share), in line with the fee income the company recognised over the decade. Loan losses never pass through operating cash flow, so we charge them separately. Measured this way, earnings were $2.10-$2.56 per share from 2016 to 2023, then fell to $1.69 in 2024 and $1.33 in 2025 as troubled loans stopped paying interest. The fall has stopped: the first half of 2025, the second half of 2025 and the first half of 2026 each produced $0.65-$0.68 per share. Book value per share fell from about $26-28 in 2016-2019 to $19.35, so we apply earning rates to today's smaller capital rather than assuming old per-share earnings return. Loan losses. From 2016 to mid-2026 the company wrote off about $1.04 billion of loans, almost all of it in 2024 ($385 million) and 2025 ($556 million), and it still holds a $398 million reserve against its loans. That is about $1.44 billion over ten and a half years, roughly $137 million a year, or about 0.8% a year of a loan book that averaged close to today's size. The cycle is not over. In the first half of 2026 the company charged $189 million to earnings for credit losses ($55 million in the first quarter, $134 million in the second), wrote off $75 million (including a $46 million loss on foreclosing a San Francisco hotel loan), and raised its reserve from $284 million to $398 million. In June 2026 a $345 million Chicago office loan made in 2018 defaulted, and three more loans totalling $502 million, two of them office and one mixed-use with office, became impaired. Office was 27% of the loan book at the end of 2025 and management puts it near 21% now. On July 30, 2026 management said it had started selling about $1 billion of mostly office loans and will market the foreclosed San Francisco Hyatt and other owned property; sale prices below the loans' carrying values would add to losses, and we charge for that in every case. Debt. Borrowing is the raw material of a lending business and interest is its main operating cost, so every value below is what is left for shareholders after lenders are paid; the borrowings are not subtracted a second time. The company reports debt of 3.9 times equity, says 88% of its borrowings cannot be called on changes in market value, and had $1.2 billion of liquidity. The 3.75% senior secured notes due January 2027 ($335 million) and the 5.50% convertible notes due March 2027 ($266 million) were pre-funded in May 2026 by $450 million of 6.25% senior secured notes due 2031, which adds only about $7 million a year of cost; the next corporate maturities are the B-7 term loan and 7.75% notes in 2029. The constraint that matters is a covenant in the company's credit agreements: tangible net worth, as the lenders define it, must stay above $3.6 billion plus 75%-85% of the cash raised from any new shares; earnings must cover fixed charges 1.3 times. The company does not publish the lenders' definition. Our judgement is that it adds the loss reserves back to book equity (equity alone was already below $3.6 billion at the end of 2025, when the company was in compliance). On that reading the cushion was about $195 million at the end of 2025 and about $70 million at June 30, 2026, because $158 million of dividends were paid in the half against a $88 million loss. Because new shares raise the minimum by most of what they bring in, issuing shares barely helps; the workable responses are a lower dividend, a smaller loan book, asset sales, or lender amendments, and one lender already lowered its minimum to $3.0 billion in 2025. The dividend of $1.88 a year (about $317 million) is well above what the business earns after losses, and management told investors on July 30, 2026 that it would review long-term earning power with the board. A lower dividend keeps capital inside the company; it does not destroy value. Base case, about $15.80 per share (about $2.7 billion for all shareholders after lenders). Year one: earnings before losses hold at the current $1.33 per share plus fees, about $1.69 (8.75% on equity). This is flat against the 21% fall in 2025, and we hold it flat rather than lower because the fall ended in early 2025 and three straight half-years have been level. Credit costs are $250 million in year one and $175 million in year two, below the first-half 2026 pace but well above normal, as the office loans are sold and the Chicago and other impaired loans are resolved. From year three, earnings before losses return to about 9.7% of equity, the company's 2016-2025 median, as roughly $2 billion of non-earning loans and foreclosed property are turned back into paying loans, as it did when it resolved eight impaired loans in 2025. Losses then run at the 0.8% through-cycle rate. That leaves about $1.01 per share a year, roughly 5.2% on book value, close to the roughly 5.4% a year the company actually earned on its capital, after losses, over 2016-2025. The dividend is reset to what earnings support, and lenders adjust the net-worth test if needed, as one did in 2025. We credit no growth in the loan book. Bear case, about $6 per share (about $1.0 billion today). This case includes one full downturn of the kind commercial property lenders have actually been through. First, the current office cycle runs on at the first-half 2026 pace: credit costs of $380 million in each of the next two years and $250 million in the third, about $1.0 billion in all, wiping out earnings, ending the dividend, and pushing book value down to about $17.60 per share. Lenders demand amendments and paydowns, costing about $25 million a year in fees and higher spreads. Then a broader crash like 2008-2010, when US commercial property values fell roughly 35%-40%, hits in years five to seven. A senior loan at 65% of value on a property that falls 40% loses about a sixth to a third of its principal once foreclosure costs are counted; we assume losses equal to 9% of the loan book over three years, about $1.3 billion, half again the roughly 6% of its average book this company has charged in the narrower 2022-2026 office downturn. The company's own history shows how bad this can get: as Capital Trust it ended 2011 with shareholders' equity of negative $129 million after a March 2011 debt restructuring, although it then held riskier junior loans and bonds funded by market-priced debt. In this case lenders require new capital, and the company issues $400 million of shares at half of book value, adding about 35% to the share count. Book value ends near $11.80 per share, about 40% below today, and the business then earns about 4.25% a year on its capital. Shareholders keep something because the loans are first-ranking, 88% of borrowings cannot be called on market moves, and the company stayed within its covenants even in 2024, when it lost $204 million. Bull case, about $22.60 per share (about $3.8 billion). Reserves prove close to adequate: credit costs of $150 million and then $100 million, impaired loans and foreclosed property recycled within two years, earnings before losses back to 10.1% of equity (the median including cash fees), and losses of about 0.45% of loans a year, the low end of the company's record. After a small allowance for the manager's incentive fee, owners keep about $1.39 per share a year. We did not include a larger loan book or wider lending margins, because neither is secured. Our previous estimate was prepared on a different basis and is not directly comparable. It also treated the 2027 maturities as still to be refinanced; they were pre-funded in May 2026. Margin-of-safety price, $8. How sure are we of the earnings level? Moderately. Earnings before losses have been steady for three half-years, but the next two years of losses depend on how the office loans sell, and the long-run loss rate rests on one decade that contained one severe cycle. The worst plausible outcome is a second, broad property crash arriving before the office losses are finished while the covenant cushion is only about $70 million: the dividend would stop, lenders would force a smaller loan book, and shareholders could be heavily diluted, as in the bear case. The structure would most likely survive it, because the loans are senior, most borrowings cannot be called on market moves, and the next corporate maturities are in 2029, but the external manager is paid on equity whatever happens. If we are wrong about near-term losses (running at the first-half 2026 pace), shares are worth about $14; if earnings before losses stay at today's level, about $13; if long-run losses run at 1.0% of loans instead of 0.8%, about $12.75; and if two of these go wrong together, about $10-11. Over the holding period we estimate dividends of about $1 per share a year once the dividend is reset, with little in years one and two. From $8, if the base case plays out over five years, those payments plus a business still worth around $16 imply about 20% a year, and more if the value is recognised sooner; the bear case implies a loss of about a quarter of that price, but not all of it. The margin-of-safety price is not set higher because near-term losses cannot yet be pinned down and the covenant leaves very little room. The price also assumes the dividend has been reset to a level earnings cover, or lender amendments have restored room under the covenant. What would make this wrong: losses on the office loan sale far beyond reserves, a forced share issue far below book value, or lenders refusing to amend; on the upside, faster recycling of the roughly $2 billion of non-earning assets and losses back near the 2016-2023 record.