Adamas Trust NYMT 2026-INV5: DSCR on the Loan Tape Differed From the File on 21% of 985 Checks, Oct 2026

What was filed, and by whom

When a lender packages loans into bonds, the securitizer has to publish what any outside reviewer found when it checked the loan files. That is a Form ABS-15G under Rule 15Ga-2. Here it was signed on October 1, 2026 by HYPO Depositor 1, LLC (EDGAR CIK 2036019) for the issuing entity “NYMT Loan Trust 2026-INV5”. The reviewers call their client Adamas Trust, Inc., the company that was New York Mortgage Trust before its rebrand; the filing does not explain why the trust keeps the NYMT name. The “INV” in the name is the investor-loan series: loans on rental houses and small apartment buildings that an investor buys or refinances and repays from rent, usually called DSCR loans. Our what is a DSCR loan page explains the product, and our Constructive Loans review covers the Adamas subsidiary that makes such loans.

The filing has eight exhibits, from four reviewers, and each used its own grade key:

Source: Exhibits 99.1 to 99.7 of accession 0001104659-26-112843. Opus's grade table and pool tables are images inside Exhibit 99.3; we read them by eye and saved a transcription. Grade keys differ by reviewer, so the totals are a sum of counts, not one grade.

Two reading notes. AMC says its review “covered 100% of the securitization loan population”, yet three other reviewers cover 489 more loans in the same deal, so we read AMC's “100%” as 100% of the loans it was given. And AMC describes itself as having reviewed loans “acquired by Adamas Trust, Inc. (the “Client”) through flow reviews and acquired by reliance letter”, with origination dates from March 2022 through September 2026, so this is a bought-loan pool, not one lender's recent output.

The DSCR line: what “differed” means

A data tape is the spreadsheet of loan terms the seller gives its buyers, rating agencies and reviewers. AMC explains its test: it “compared data fields on the tape provided by the Client to the data found in the actual mortgage loan file as captured by AMC”, and for DSCR it was also “Reviewing and calculating that the DSCR and/or Property DTI conforms with applicable guidelines.” So each “difference” is a case where the number a reviewer got out of the lease, appraisal rent schedule and settlement documents was not the number on the tape.

Source: AMC's tape integrity table, Exhibit 99.1. 48 of the 64 fields listed had at least one difference. The DSCR line is 22.7% of all 915 differences (our arithmetic).

What the DSCR number tells us and what it does not:

• It is the biggest field by count. Among the fields AMC compared on more than 900 loans, DSCR has the highest share of differences (21.12%, against 9.18% for Borrower Last Name); the higher shares in the table above sit on thin samples of 27 to 64 comparisons. 985 of 1,134 loans (86.9%, our arithmetic) had a DSCR to compare; for the other 149 the filing does not say why not.
• It is not a count of bad loans. The filing gives no size for any gap. A gap can come from a lease versus an appraiser's market rent, a different tax or insurance estimate, or a keying difference. Digital Risk describes its own method as “rent per lease or comparable rent schedule from appraisal divided by the subject total PITIA payment”, and says that where lease and comparable rent differ it refers to the client's guidelines.
• Reviewers were cautious about their own output. AMC says its findings “are not statements of fact or legal conclusions”, and that it did not verify the data it relied on except as the report states.
• The DSCR gaps did not stop the grades. Every AMC loan is A or B on credit; 191 B loans carry the 231 credit exceptions, 152 of them labelled Guideline and 26 Credit, with 16 for Borrower and Mortgage Eligibility. The category list does not name DSCR, so we cannot say how many of these relate to it. Digital Risk's rebuttal report does name it: 6 of its 19 initial credit exceptions on 22 loans read “DSCR - Program Parameters - Other”.

Is this a pattern? Eight Adamas deals, same reviewer

AMC reviewed the investor loans in earlier NYMT Loan Trust INV deals too, with the same tape-integrity table. We pulled each deal's Form ABS-15G from the same depositor (CIK 2036019).

Source: Exhibit 99.1 of each ABS-15G, accessions 0001104659-24-105014, 0001104659-25-020977, 0001104659-25-090735, 0001104659-26-004889, 0001104659-26-031913, 0001104659-26-064611, 0001104659-26-084307 and 0001104659-26-112843. Pooled: 1,850 of 8,953 comparisons, 20.7%, on 9,664 loans (our sums). We left out 2024-BPL3 and 2025-CP1, which are other products with other reviewers or scopes.

Reading across: the share fell from about 28% in the first two deals to 16% in September 2025 and January 2026, and has since risen to 22.88% in July 2026 and 21.12% now. We do not know why, and the filings do not say. The grade profile moved little: A shares between 78% and 89%, no C or D in the last seven deals. For a borrower, the useful takeaway is narrow: across about 9,700 loans in eight deals over two years, a reviewer that rebuilt the DSCR from the file disagreed with the seller's tape about one time in five.

What kind of loans are in the pool

AMC's population table and Opus's pool tables (the latter in an image) agree on the broad shape:

Source: Exhibits 99.1 and 99.3 (Opus tables are images, transcribed by us). Opus also lists 36 loans (8.20%) as “Other” property type, and says 63 borrowers have more than one loan in the pool.

Infinity's 28 loans are the only group where the originator is named: MoFin Lending Corporation. 12 of them have a 240-month amortization and 16 a 360-month one, which matters for the DSCR: a shorter amortization means a larger monthly payment for the same balance.

Adamas Trust and Constructive: what the 10-Q adds

Adamas Trust's Form 10-Q for September 30, 2026 (accession 0001273685-26-000078, filed October 30, 2026) says that “through our wholly-owned subsidiary, Constructive Loans, LLC”, it originates business purpose loans for real estate investors. In its held-for-sale roll-forward, the principal balance of loans originated was $376,458 thousand in the third quarter and $1,183,397 thousand in the nine months (our reading: $376.5 million and $1,183.4 million). Of the nine-month figure, $404,509 thousand were sold to third parties and $709,329 thousand were transferred out of held for sale to Adamas's own residential loans, and $147,591 thousand was left at September 30, 2026.

The 10-Q's rental-loan book is another useful yardstick for a borrower: 13,247 business purpose rental loans, $2,617,300 thousand unpaid principal, weighted average FICO 751, loan-to-value 70% and a weighted average coupon of 7.01%. The 10-Q says one business purpose rental loan securitization closed in the quarter, for about $338.9 million of net proceeds; it does not name the trust, so we do not link it to INV5.

Unverified link. The INV5 exhibits do not name an originator for AMC's or Opus's loans, only for Infinity's. Some of the loans may be Constructive's, but nothing in the filing says so. See our Constructive Loans review for what Constructive publishes about its own programs, and our BPL Mortgage review for a large separate DSCR-type lender.

What a reader can do with this

1. Ask for the DSCR worksheet and the rent source. The tape-versus-file gaps in this deal were found by reviewers who rebuilt the number from the lease or appraisal. Ask your lender which rent it used (lease or market rent), the PITIA it divided by, and get both in writing.
2. Rerun your own ratio. Use our DSCR loan calculator with your actual tax and insurance bills. Illustration (our arithmetic, not from the filing): rent of $2,100 a month over PITIA of $2,000 is 1.05x; if taxes and insurance are really $150 higher, PITIA is $2,150 and the ratio is 0.98x, below 1.00x. A lender with a 1.00x minimum would then decline or reprice.
3. Watch PITIA and prepayment terms, not just the ratio. AMC's tape differed on Original PITIA in 24 of 54 comparisons and on Prepayment Penalty in 19 of 27 (small samples). Check the note against your term sheet before closing.
4. Compare minimums before you apply. Lenders set different floors; see DSCR loan requirements in 2026 and the ratio rules in DSCR minimums for Fannie, SBA, HUD and rental lenders.
5. Do not read this as a verdict on any lender. For another reviewer's grades on short-term loans, see our page on Vontive's 2026-RTL1 diligence and the Verus Securitization Trusts page.

This is analysis of public documents, not investment, legal or tax advice. It does not say that Adamas Trust or any originator is a better or worse lender than another.

Kiavi pays us a referral fee when a loan closes through the button in the box below. The counts and grades above are the same either way, and Adamas Trust and Constructive are not affiliates. Visio and Lima One are shown as alternatives.

FAQ

Figures are from Form ABS-15G accession 0001104659-26-112843 (HYPO Depositor 1, LLC, CIK 2036019, dated October 1, 2026, Exhibits 99.1 to 99.8 from AMC Diligence, Opus, Digital Risk and Infinity), the seven earlier INV-series ABS-15G accessions listed under the series table, and Adamas Trust, Inc.'s Form 10-Q accession 0001273685-26-000078 filed October 30, 2026, all read on November 4, 2026. Counts and shares marked as ours were computed by CrowdfundedWealth from the saved exhibits. Opus's tables are images inside Exhibit 99.3 and were transcribed by us. Grades and discrepancy findings are as stated by the reviewers.