What happened. Almost every headline put this week’s Toorak deal at $3.2 billion.4 Velocity Financial’s own 8-K puts the price it is actually paying at approximately $62 million.1 Both numbers are real, and the gap between them is the story. Velocity agreed on August 26 to buy the operating platform of Toorak Capital, the KKR-backed business-purpose lender, while an entirely different buyer takes the loans.
Velocity Commercial Capital is buying the platform. A separate, unrelated buyer is taking Toorak’s existing book of roughly $3 billion in unpaid principal balance. The $3.2 billion figure is the two transactions added together, valued off Toorak’s June 30 balance sheet.2
The press release calls that buyer “a third-party investment firm” and never names it. The 8-K does: affiliates of TOBI III SPE I LLC, with a related agreement signed with Titan RR LLC.1
Velocity’s purchase is conditioned on that separate sale closing. If the back book does not trade, the platform deal does not close either.1
What Alternative Business Lenders Need to Know
This is a residential transition loan deal, and your book is not RTL. The duration is different, the loss curve is different, and your securitization exit, if you run one, prices off different collateral. None of that transfers.
The structure transfers, and you have already seen it in your own asset class. In July we covered Fortress agreeing to buy up to $1.5 billion of Wayflyer’s merchant advances over three years, a forward flow that leaves Wayflyer originating and servicing paper it does not hold.19 That is the same split Velocity just paid for. Someone else owns the receivable; the originator keeps the origination fee and the servicing strip. Every MCA funder syndicating a share of each deal to participants, every factor selling participations, and every equipment shop discounting paper to a funding source is already running a version of it. What is unusual here is not the structure. It is that the two halves were priced separately, in public, by different buyers, which is the number you normally never get to see.
What did Velocity buy, and who took the loans?
The people, the technology, the channels, and the right to keep originating. The 8-K describes a purchase of equity interests in companies carved out of Toorak Capital Partners LLC, MMTC Holdings LLC and MMTC Portfolio Holdings LLC through a statutory division, for “a base purchase price of approximately $62 million, plus estimated tangible book value, subject to certain adjustments.”1 Velocity’s own investor deck rounds it to $63 million, which is the figure the trade press picked up.3 5
What that buys is substantial. Toorak has funded more than $20.3 billion since 2016 across roughly 43,000 loans, produced $2.9 billion in FY 2025 volume, and employs about 280 people, including roughly 120 at Merchants Mortgage and Trust, its direct origination arm.3 2 Toorak keeps its brands, its management team, and its Tampa headquarters, and founder John Beacham becomes an executive vice president of Velocity Commercial Capital.2 Velocity says the deal lifts its production volume from about $2.7 billion to about $4.8 billion, a 76% increase, and expands servicing by 39%.3 Roughly $62 million for an origination engine that nearly doubles your volume is not an obviously expensive trade.
This is the part worth reading carefully, because the press release and the filing say different amounts. Exhibit 99.1 refers only to “a third-party investment firm.”2 The 8-K identifies the counterparty as affiliates of TOBI III SPE I LLC, and discloses a Master Transaction Agreement between Velocity Commercial Capital and Titan RR LLC, described as an affiliate of an affiliate of TOBI, under which TOBI will pay Velocity certain securitization-related fees after closing.1
The book itself is roughly $3 billion of unpaid principal balance across whole loans and positions held in Toorak Mortgage Trust and TRK Trust securitizations.2 Velocity does not buy it. Velocity will manage it, and separately agreed to sell future Toorak production to the same firm and other counterparties.2 The company puts combined assets under management at roughly $10 billion of unpaid principal balance after closing.3
Why split the platform from the book at all?
Because they are two different businesses with two different costs of capital, and the market prices them very differently. Holding $3 billion of short-duration loans requires equity, warehouse capacity, and a tolerance for credit losses. Originating and servicing those same loans earns fees and requires almost none of that. Velocity is explicit that the point is “a capital-light, high-return business” whose earnings come from “origination-related fees, servicing / asset management fees” supported by forward loan sale agreements.2
A firm with a decade of performance data on this asset class, backed by KKR, sold the loans and kept none of the credit risk, while a public competitor paid to own the machine that makes them. Read that as pricing rather than prophecy. What is observable is narrow and still useful: two halves of one lender were valued separately, by different buyers, at separately negotiated numbers.
Is the market growing into this deal or out of it?
Out of it, at least on the securitization side, which is what makes the timing interesting. KBRA published research on July 30 projecting full-year 2026 RTL securitization issuance of approximately $4.8 billion, down from $5.5 billion in 2025 and $6.2 billion in 2024, with rated issuance around $3.5 billion, or roughly 73% of the projected total.6 KBRA also notes RTL remains under 5% of expected 2026 private-label securities issuance.6
So Velocity is paying to nearly double origination in a product whose securitization exit is forecast to shrink about 12% this year. That is not necessarily a mistake. Rated deals are taking share of a smaller pool, and a firm selling loans forward to a named buyer is less exposed to the public market than one that has to print a deal every quarter. But it does mean the fee stream depends on forward flow appetite holding up.
What does Toorak’s credit record actually show?
Velocity’s deck reports cumulative net principal losses of less than 10 basis points across Toorak’s lifetime U.S. and U.K. single-family and multifamily transitional lending, and a weighted average life of about 15 months on paid-off transitional loans.3 Fewer than 10 basis points of cumulative losses on more than $20 billion of lifetime volume is a genuinely strong number, and it is the best argument for the price.
Two honest caveats. This figure comes from the buyer’s own marketing deck, not from an audited statement or a rating agency surveillance report, and the footnote defines it narrowly as transitional loan activity, which excludes the DSCR rental product. And a 15-month asset that has mostly been originated into a rising housing market has not been tested the way a longer-duration book has. Short duration is a real risk control, because the book turns over before a downturn compounds. It is not the same thing as proof of underwriting quality across a cycle.
Has Velocity been here before?
Yes, and its own filings are blunt about how the last shock went. Velocity priced its IPO at $13.00 per share on January 17, 2020, below its $14 to $16 marketed range.7 9 Ten weeks later the pandemic hit. Its FY 2020 10-K states that the company “temporarily suspended our loan originations and loan purchases from late March through August and furloughed a significant number of our employees, mostly within our loan origination function,” resuming in September.8
That history is the strongest argument for what it is doing now. A lender whose originations went to zero for five months because its funding and its balance sheet were the same thing has an obvious reason to want fee income that does not require holding the asset. Velocity expects the transaction to be accretive to GAAP earnings in 2027 and estimates book value dilution of 4% to 6%, earned back in about three years.3 Those are management projections, not results.
What should lenders do differently this week?
Separate your two profit pools on paper and look at them honestly. Calculate what your origination and servicing operation earns as a standalone fee business, with a market-rate charge for the capital your balance sheet provides, and calculate what your held assets earn net of funding, losses, and the equity they tie up. You have almost certainly modeled both. What you have probably not done is price them as two businesses that could be sold to different buyers on the same afternoon, which is exactly what happened here.
Then stress the forward flow assumption. If your growth plan quietly depends on selling production to a small number of buyers, write down what happens to originations in the quarter after your largest buyer pauses. Velocity’s platform purchase does not close unless the back book sale closes, which is a useful reminder that in a capital-light model your counterparty’s appetite is a closing condition, not a detail.
Our Opinion
The reporting on this deal has mostly been a number, $3.2 billion, attached to the wrong noun. Velocity is not buying a $3.2 billion company. It is paying roughly $62 million for an origination and servicing platform while a different firm absorbs $3 billion of loans, and the two events are stapled together because neither closes without the other. Once you see it that way, the deal reads less like consolidation and more like an unbundling.
The tempting read is that KKR, after a decade in this asset class, took the exit on the credit because it knows something. We would not lean on that. A fund reaching the end of its life, or simply a strong bid on a clean book, explains the same facts without implying any view on the collateral. The part that does not need a motive is the pricing. Velocity, which had its originations shut off for five months in 2020, paid for the half that earns fees without requiring capital, and a different buyer paid for the half that carries the risk. For an industry where most operators still hold both halves by default and never see them quoted apart, that is the useful disclosure.
The caution we would attach is that capital-light is not risk-light, it is a different risk wearing a friendlier name. Toorak’s sub-10 basis point loss record is impressive and it is also, for Velocity, no longer the thing that determines earnings. The number that matters now is whether TOBI and the other counterparties keep buying, and that is not a number anyone publishes.
1-Minute Video: Construction Lending Verification: Most Lenders Stop at Layer One
Most construction lending files stop at one check: the business exists. That confirms registration, not authorization. A company can hold a clean Secretary of State status and still have a lapsed or missing contractor license for the work it is about to be funded to perform.
Cobalt walks through the three layers it verifies: a Secretary of State check that confirms the business legally exists, a Contractor License API check that confirms the contractor is actually authorized, and a court records search that surfaces litigation a credit review would miss. Most lenders stop at layer one.
Free Tools for Lending Teams
Headlines You Don’t Want to Miss
Colorado Attorney General Phil Weiser sued Activehours, Inc., doing business as EarnIn, on August 27. The complaint alleges the company’s “Cash Out” product is an illegal high-cost payday loan, and that between January 2023 and July 2025 EarnIn made more than 3.1 million loans to 56,778 Colorado consumers, lending roughly $300 million and collecting more than $16 million in tips and Lightning Speed fees, at an average APR near 388% and in some cases above 1,000%.10 11 These are allegations in a complaint, not findings by any court, and the state says the suit targets only the direct-to-consumer product, not employer-integrated wage access.10 Before you read this as a trend, note the counterweight: a D.C. trial court dismissed most of the District’s parallel claims against EarnIn, and in February 2026 the D.C. Court of Appeals left that dismissal in place, on reasoning that whether earned wage access is a loan belongs with the financial regulator rather than an enforcement suit.12 The fee-versus-finance-charge question is being litigated state by state, and it is the same question that decides whether your own expedite fees and optional tips are disclosed as a cost of credit.
The SBA published two proposed rules on August 20 that would rewrite how a business qualifies as small, consolidating the size standards and shifting many industries onto employee-count tests.13 14 16 The agency estimates the change would make roughly 114,500 additional firms eligible as small businesses. Reaction has split along predictable lines: lenders see a larger pool of established, bankable borrowers for 7(a) and 504 paper, while smaller firms argue they will be competing against companies several times their size for the same set-asides.15 The practical point for non-bank lenders is that the borrower you lose to a government-guaranteed product may be a larger borrower than it used to be. Comments on both rules close September 21, 2026, which is a real deadline and a short one.13
The FDIC closed Tioga-Franklin Savings Bank of Philadelphia on August 21, its fifth failure of 2026, with about $68 million in assets and $67 million in deposits transferred to Second Federal Savings and Loan Association of Philadelphia at an estimated cost of $5.5 million to the Deposit Insurance Fund.17 18 Five failures against two in each of 2024 and 2025 is an uptick, though all five were small.17 The detail worth noticing is on the same FDIC list: the July 17 failure was Small Business Bank of Lenexa, Kansas.17 Community institutions are the incumbent competitor for exactly the borrower most alt-lenders underwrite, and each closure hands a book of small-business relationships to an acquirer that may not want all of them. That is a referral pipeline, not just a headline about bank health.
Sources
1 SEC EDGAR | Velocity Financial, Inc. Form 8-K, Items 1.01, 7.01, 9.01 (filed August 27, 2026; event dated August 26, 2026)
2 SEC EDGAR | Velocity Financial, Exhibit 99.1, Press Release (August 27, 2026)
3 SEC EDGAR | Velocity Financial, Exhibit 99.2, Toorak Capital Transaction Overview investor presentation (August 2026)
4 Alternative Credit Investor | Velocity snaps up RE lender Toorak and $3bn portfolio (August 27, 2026)
5 Scotsman Guide | “Godfather” of RTL securitizations breaks down Velocity’s $63M Toorak deal (August 27, 2026)
6 KBRA | Laying the Foundation: The Evolution of RTL Lending (July 30, 2026)
7 SEC EDGAR | Velocity Financial, Inc. Form 424B4 initial public offering prospectus (January 17, 2020)
8 SEC EDGAR | Velocity Financial, Inc. Form 10-K for fiscal year 2020 (filed March 17, 2021)
9 Renaissance Capital | Velocity Financial prices IPO at $13, below the range (January 16, 2020)
10 Colorado Attorney General | Weiser sues EarnIn for illegal high-cost lending, deception, and violating state payday lending and consumer credit laws (August 27, 2026)
11 Law360 | Colo. AG Accuses EarnIn Of Making Illegal Payday Loans
12 American Banker | D.C. appeals court hands favorable ruling to EWA fintech EarnIn
13 Federal Register | Small Business Size Standards, Proposed Rule, 2026-17042 (published August 20, 2026; comments close September 21, 2026)
14 Federal Register | Small Business Size Standards: Revised Size Standards Methodology, Proposed Rule, 2026-17039 (August 20, 2026)
15 American Banker | Small firms hate SBA’s new size standards; lenders are intrigued
16 SBA Office of Advocacy | SBA Issues Proposed Rules on Industry Size Standards and Revised Size Standards Methodology (August 20, 2026)
17 FDIC | Failed Bank List
18 Banking Dive | Philly lender becomes fifth bank to fail in 2026
19 Beyond Banks | Fortress Will Buy $1.5B of Merchant Advances (July 30, 2026)
