Cherry's $350M Securitization Isn't the Deal - It's the Infrastructure Proof Point

KBRA assigned preliminary ratings to four classes of notes in Cherry Securitization Trust 2026-1 on June 1, totaling $350 million - collateralized by receivables for elective medical procedures. Cherry, a private fintech that provides patient financing for dental, veterinary, plastic surgery, and medspa services, now has three consecutive securitization deals on the books: $400 million in 2024 (up from an initial $250 million mooted size), $300 million in 2025, and this $350 million tranche.

The deal itself is not the signal. The signal is what the securitization pipeline proves: machine-learning-driven credit underwriting has moved from prototype to institutional infrastructure, and alternative markets - not just traditional personal loans - are now bankable through structured credit vehicles.

This is the product architecture transition this persona watches. The market is shifting from FICO-dependent credit decisions to ML-underwritten underwriting, and securitization is the metric that tells you whether institutional capital has caught up.

The architecture shift: from FICO gates to ML approval

Cherry's product architecture is built on a different underwriting stack than the incumbent consumer lenders. No hard credit checks - only soft pulls that don't touch credit scores. Approximately 90% approval rate across credit profiles. True 0% APR financing. Applications approved in 35 seconds. Up to $50,000 in funding, paid upfront to healthcare providers while patients repay over time.

Most of that approval rate - the part that separates Cherry from a traditional lender - comes from underwriting models that process data points FICO simply doesn't capture. Income stability, treatment type, provider relationship, repayment behavior across alternative credit sources. That's the ML architecture advantage: more variables, faster decisions, broader access.

The debate is not whether AI underwriting works. The debate is whether it works well enough at scale to satisfy institutional investors through a securitization - where the ratings agency has to model loss scenarios and the investors have to trust the credit enhancement levels are real.

Cherry's pipeline says yes. KBRA has rated three consecutive tranches and affirmed ratings on the earlier trusts in May 2026. That's institutional surveillance, not a one-off.

Deal size as the leading indicator

Look at the deal progression:

The 2024 deal is the most telling data point. Barclays served as structuring lead, and the deal was upsized from $250 million to $400 million on the back of strong investor demand. That's not a company asking for credit - that's investors queuing up for an asset class they didn't believe in two years prior.

Cherry's loan originations grew 150% year-over-year heading into 2025. The company raised $94 million in total funding through September 2022 across eight rounds, including a $50 million debt financing. Since then, securitization has replaced venture capital as the primary funding engine. That is the hardware-to-software value migration pattern: once the underwriting engine - the software - proves itself, the balance sheet structure shifts away from equity risk and into structured debt. The underwriting technology becomes the asset that generates the funding.

What this means for the publicly traded AI lending trade

Here's where the capital allocation question kicks in, because Cherry is private - you can't buy the company. The question is whether Cherry's trajectory tells you something about the publicly traded competitors who are running the same underwriting architecture.

Upstart, the closest public analog, just completed its 50th ABS securitization in April 2026, with KBRA also providing preliminary ratings. The difference is scale and transparency: Upstart's securitization history is visible, priced, and subject to earnings calls. Cherry's is a private proof point that validates the broader thesis - AI underwriting is infrastructure now, not a niche experiment.

The market structure transition runs like this: traditional lenders built their moat on FICO and relationship data. AI-native lenders are building theirs on ML models that process broader data, approve more borrowers, and securitize the resulting pools at competitive ratings. The moat isn't the model itself - it's the data flywheel. More approvals generate more repayment data, which improves the model, which justifies more securitization capacity, which funds more originations.

This is what separates companies on the right side of this transition from those still running legacy credit infrastructure.

The risk that matters

However, there's a structural risk in this architecture that the deal press release doesn't address. Cherry's collateral pool - elective medical procedures - is concentrated in discretionary healthcare spending. Dental, cosmetic surgery, medspa treatments. These are the first consumer expenses to compress when the employment picture weakens. A FICO gate exists for a reason: it screens for exactly this type of downside correlation.

ML underwriting can widen the approval pool, but it can't change the fundamental question of whether elective medical receivables hold up in a recession. Cherry's 90% approval rate is impressive in a stable economy. I don't have the data to tell you how those portfolios behave when borrowers are choosing between a dental crown and a car payment.

The previous KBRA rating affirmation in May 2026 covers only current performance - not stress scenarios. That's a data gap, and it matters if you're evaluating the durability of this underwriting thesis.

Where the capital goes

Cherry can't be bought. But the signal is actionable. The question is whether you believe AI-driven credit underwriting has moved from innovation story to baseline infrastructure - or whether it's still a niche bet with untested downside behavior.

I believe the securitization pipeline proves the infrastructure case. Three consecutive rated tranches, a 150% origination growth rate, and the shift from venture funding to structured debt - that's not a company asking for patience. That's a company that has cleared the institutional adoption barrier.

For portfolio allocation, the publicly traded proxy is Upstart. Its 50th securitization and the same ratings partner (KBRA) put it on the same institutional track, with the advantage of earnings visibility and price discovery. If Cherry's pipeline validates the thesis, Upstart's stock price is where that thesis gets tested in real time.

The debate is not whether AI underwriting works. It is whether the return profile of publicly traded AI lenders still justifies the allocation - especially when the private players like Cherry are proving the model without the earnings pressure that comes with being public. That's the opportunity cost question.

Cherry's $350 million securitization isn't a deal to track. It's an infrastructure proof point. The capital should go where the same architecture trades with transparency, earnings cadence, and enough volatility to separate the signal from the noise.