Company Research • Initiating coverage • October 30, 2026 KST / October 29 US. Nasdaq: PGY. Financial results through September 30, 2026; valuation horizon October 2027 using FY2027 earnings. All amounts are US dollars. This is our first PGY fundamental model, not a revision of an earlier AHNIM target.
Can Pagaya turn faster loan-network growth into durable per-share earnings? Its third-quarter results show operating leverage, but not yet improving unit economics: network volume rose 28.7% while fees less production costs grew only 5.1%. Our separate cash bridge also finds that nine-month operating cash flow did not cover capital expenditure and net purchases of credit investments. We initiate conditional October 2027 values of $7.48 Bear, $22.75 Base and $45.03 Bull. The Base offers only 9.3% price upside to the October 29 regular-session close of $20.81; it requires better fee retention without assuming away investment losses or dilution.
The new evidence in this analysis is the connection between three measures that should not be conflated: profit earned on each dollar of network volume, accounting earnings after investment marks, and cash remaining after retained credit investments. Our valuation keeps those costs visible instead of applying a software multiple to adjusted EBITDA.
1. The business: technology distribution with a credit-funding requirement
Pagaya connects lending partners with institutional funding for consumer credit. Its technology helps partners evaluate and monetize loan applications, while financing vehicles acquire the resulting assets. The economic product is not simply an AI subscription: the network needs borrowers, willing lending partners and investors willing to fund credit at an acceptable price. Pagaya earns network AI fees, including integration and capital-markets execution, plus contract fees such as administration, management and servicing. Its June 2026 Form 10-Q distinguishes point-in-time network fees from continuing contract services.
Personal loans, auto loans and point-of-sale financing bring different borrower and collateral exposures. Partners can use their own underwriting and balance sheets, work with alternative funding channels, or change which applications enter the network. Our assessment of the competitive advantage therefore rests on incremental approvals that remain profitable after losses and funding costs—not on application volume or AI branding alone. A larger network can improve distribution and data coverage, but those benefits are not a guarantee against adverse borrower selection or a credit-cycle turn.
The October 29 shareholder letter names Citizens Financial Group for personal loans and FNBO for home-improvement point-of-sale lending as new partners. We treat those relationships as distribution opportunities, not as a disclosed revenue backlog. Partner onboarding, funded volume and profitable cash generation are separate milestones. Announced securitizations and forward-flow arrangements are funding capacity or asset transfers, not revenue; overlapping funding announcements should not be added into a single incremental sales figure.
2. Q3: volume accelerated, but the retained fee yield fell
GAAP means generally accepted accounting principles; EPS means earnings per share. EBITDA is earnings before interest, taxes, depreciation and amortization, with the company making additional adjustments in its reported non-GAAP measure. A basis point is one hundredth of a percentage point.
| Metric | Q3 2026 | Q3 2025 | AHNIM reading |
|---|---|---|---|
| Network volume | $3,605m | $2,802m | +28.7% |
| Fee revenue | $362.155m | $339.887m | +6.6% |
| Production costs | $215.701m | $200.574m | Paid before retained fee economics |
| Fee revenue less production costs (FRLPC) | $146.454m | $139.313m | +5.1% |
| FRLPC / network volume | 4.063% | 4.972% | Down 90.9 basis points |
| Total revenue and other income | $390.492m | $350.165m | +11.5% |
| GAAP operating income | $112.548m | $79.958m | +40.8% |
| Net income attributable to Pagaya | $49.732m | $22.545m | +120.6% |
| GAAP diluted EPS | $0.52 | $0.23 | Reported EPS; not net income / basic shares |
FRLPC equals fee revenue minus production costs. Dividing it by network volume gives a useful retained-fee yield, not a net margin or a return on the credit assets. The calculation is $146.454m / $3,605m = 4.0625%, versus $139.313m / $2,802m = 4.9719%. More loans therefore did not translate proportionately into more retained fees. This matters because a valuation built only on volume growth would miss the economic compression.
Operating profit nevertheless improved because expenses below production costs fell: technology, sales and marketing, and general and administrative costs totaled $62.243m, versus $69.633m. Interest income also increased to $28.962m from $14.918m. The resulting operating leverage is real in the reported accounts, but the interest income comes with investment and funding exposures that must remain in the analysis.
Management attributes pressure on retained fees to funding conditions and expects consumer pricing actions to help subsequently. It guides Q4 FRLPC toward the lower end of its 4%–5% range and expects improvement toward the middle of that range in 2027. These are management expectations, not already realized margins. Our Base uses 4.3%, below the 4.5% midpoint: partial recovery rather than full normalization.
3. Why adjusted EBITDA is not the shareholder earnings base
Q3 adjusted EBITDA was $130.252m, compared with $49.732m of net income attributable to Pagaya. The reconciliation adds back, among other items, $36.932m of impairment on certain investments, $8.607m of stock-based compensation and $22.483m of interest expense. Those adjustments help explain the company measure; they do not make credit exposure, compensation or financing free.
The income statement separately records $43.400m of losses on investments in loans and securities below operating income in Q3, and $123.714m over nine months. These are investment and fair-value losses, not necessarily cash defaults in the same period. They are also distinct from the smaller investment-loss line included in total revenue and other income. We do not subtract either line twice. Our forward model retains an explicit allowance for below-operating investment losses and includes compensation within operating expenses.
Net income and diluted EPS require different denominators and, for convertible securities, different numerators. Q3 diluted weighted-average shares were 98.474m, versus 83.673m basic shares. Dividing $49.732m by a quote-feed share count is not a reconstruction of GAAP diluted EPS. Preferred-share participation and the if-converted treatment of exchangeable notes matter.
4. The cash bridge: profitable does not mean self-funding
| Cash-flow item ($m) | 9M 2026 | 9M 2025 |
|---|---|---|
| Operating cash flow | 166.937 | 158.823 |
| Capital expenditure: PP&E and software | (9.907) | (10.816) |
| Operating cash flow less that capex | 157.030 | 148.007 |
| Purchases of loans and securities | (782.989) | (361.070) |
| Maturities and prepayments | 510.790 | 146.809 |
| Sales of loans and securities | 19.394 | 31.387 |
| Net cash used for those credit investments | (252.805) | (182.874) |
| Selected residual after capex and net credit investments | (95.775) | (34.867) |
The formula is operating cash flow − capital expenditure − [investment purchases − maturities/prepayments − investment sales]. For 2026: $166.937m − $9.907m − [$782.989m − $510.790m − $19.394m] = −$95.775m. Positive operating cash flow is therefore not evidence that growth financed itself after these retained investments.
This does not mean the company lost $95.775m in economic value. Investment purchases exchange cash for assets; proceeds can arrive later. It does mean shareholders must examine the financing and subsequent realizations. The full statement shows $263.198m of net investing outflow, $112.330m of financing inflow and $1.142m of currency effects, which together with operating cash flow reconcile to a $17.211m increase in cash and restricted cash. Our selected residual excludes other investing flows and is not a distributable-cash measure.
At September 30, unrestricted cash was $247.115m and restricted cash $58.445m. Loans and securities at fair value were $1,100.832m. Revolver, secured borrowing, exchangeable-note and long-term-debt carrying values summed to $962.380m. Those assets and liabilities are central to the earnings engine. We neither treat the entire investment portfolio as spare cash nor subtract all financing debt again from an equity P/E valuation that already includes financing expense.
5. What the next quarter needs to confirm
The October 29 full-year outlook is network volume of $12.725bn–$13.025bn, total revenue and other income of $1.450bn–$1.500bn, adjusted EBITDA of $465m–$485m and GAAP net income of $162m–$182m. Against reported nine-month results, the implied Q4 net-income range is approximately $42m–$62m. The range is not a promise, and the financial statements remain unaudited. We use the Q3 earnings package for current results and the Q2 10-Q for the older detailed financing and equity notes; we do not present June note disclosures as newly filed September details.
The sequence to watch is partner activation → funded volume → retained fee yield → investment realizations → cash available after financing needs. A fourth-quarter rise in volume with another fall in yield would not establish the Base case. Conversely, stronger retained yield and lower losses could improve earnings without requiring another unusually large volume increase.
6. October 2027 valuation: retain credit costs, then price the earnings
Our primary method is FY2027 modeled earnings per economic share multiplied by an assumed P/E. It is a forward scenario value, not a discounted present value or a precise GAAP EPS forecast. The core pre-tax formula is network volume × FRLPC yield + other revenue/income − GAAP-style operating expenses − below-operating investment losses − other expense including financing. Stock compensation and ordinary depreciation remain within expenses. The other-income input is net of the investment items included above operating income; no separate add-back is made for them.
| AHNIM FY2027 assumptions ($m unless stated) | Bear | Base | Bull |
|---|---|---|---|
| Network volume ($bn) | 13.5 | 15.0 | 17.0 |
| FRLPC yield | 4.0% | 4.3% | 4.6% |
| FRLPC | 540 | 645 | 782 |
| Other revenue and income | 100 | 115 | 130 |
| Operating expenses after production costs | 270 | 270 | 285 |
| Below-operating investment losses | 180 | 160 | 145 |
| Other expense including financing | 95 | 95 | 90 |
| Pre-tax income | 95 | 235 | 392 |
| Normalized tax assumption | 15% | 15% | 15% |
| Modeled net income before preferred allocation | 80.75 | 199.75 | 333.20 |
| Preferred participation allocation | 3 | 5 | 7 |
| After-tax exchangeable interest add-back | 0 | 10 | 10 |
| Earnings numerator | 77.75 | 204.75 | 336.20 |
| Economic share denominator (m) | 93.5653 | 108 | 112 |
| Modeled earnings per share ($) | 0.8310 | 1.8958 | 3.0018 |
| Assumed P/E | 9× | 12× | 15× |
| October 2027 value ($/share) | 7.48 | 22.75 | 45.03 |
Base arithmetic: ($15,000m × 4.3% + $115m − $270m − $160m − $95m) × 85% = $199.75m. After $5m of assumed preferred participation and a $10m after-tax if-converted interest adjustment, $204.75m / 108m shares × 12 = $22.75. We assume zero net non-controlling-interest attribution in the forward bridge. The normalized 15% tax rate is an analytical allowance rather than a forecast statutory rate; geography, deferred taxes and realizability can change the outcome.
The Base combines network growth with only partial fee-yield recovery. Its $270m operating-expense allowance is above annualized Q3 costs, while its $160m investment-loss allowance is close to annualized nine-month losses, rather than zero. Other income and financing expenses scale from the current interest-bearing business. The Bull needs higher yield and a lower absolute loss burden despite more volume; that is a demanding credit-and-funding outcome, not simply a stronger AI adoption story. The Bear combines weaker economics with a lower valuation multiple.
The 9×/12×/15× multiples are our judgments, not asserted peer medians. The Base recognizes a profitable distribution platform but discounts funding dependence, volatile marks and dilution. A premium would require evidence that cash generation and losses remain resilient across more than one favorable quarter. We do not add cash or deduct debt a second time, and we do not assign a DCF value from operating cash flow minus capex while ignoring required credit investments.
7. Dilution is part of the model, not a footnote
The June 10-Q reported $160m principal of 2029 exchangeable notes, a 6.125% coupon and an approximately $13.99 conversion price. Its Q2 diluted EPS calculation included about 11.435m if-converted shares and an associated interest adjustment. Our Base and Bull denominators already include conversion exposure, so those shares are not added again. Their $10m after-tax interest add-back is a rounded modeling allowance, not the exact future accounting adjustment.
June disclosures also showed 19.883m options to restricted shares excluded from diluted EPS as anti-dilutive, with a $19.30 weighted-average exercise price. These do not all become incremental shares one-for-one at every stock price: exercise proceeds, vesting and treasury-stock treatment matter. Our Base 108m and Bull 112m denominators allow for additional option/award exposure and future dilution above the Q3 98.474m diluted average. They are conservative scenario assumptions, not a verified fully diluted current share count or a formal option-by-option treasury-stock calculation. We do not separately add assumed exercise proceeds to equity value.
In the Bear, including the modeled $10m convertible interest benefit would be anti-dilutive at the assumed earnings level. We therefore remove both that add-back and 11.434704m conversion shares from a 105m economic-claim envelope, leaving 93.565296m shares. The debt and its financing cost remain. This prevents a mechanical conversion assumption from increasing low-case earnings per share.
| One change to Base; all else held constant | Effect on modeled value |
|---|---|
| FRLPC yield +/− 0.50 percentage points | +/− $7.08 per share |
| Investment losses $50m higher | −$4.72 per share |
| Share denominator 112m instead of 108m | $21.94 per share |
| P/E 10× instead of 12× | $18.96 per share |
8. Market benchmark and the 21.1% analyst-target gap
The October 29 regular-session quote was $20.81 at 4:00 p.m. EDT, with a vendor-reported market capitalization of approximately $1.74bn. This remains below the historical $10bn small-cap screening reference. Quote-feed outstanding shares are not our forward diluted denominator, and after-hours prices are not substituted for the completed close.
The accessible StockAnalysis/S&P Global target snapshot lists 11 analysts, a $28.82 mean, $27 median and $23–$36 range for rolling 12-month targets. Crucially, the displayed target update is October 6, with a last-check date of October 20: it is a pre-Q3-results benchmark, not a verified post-earnings consensus. Its approximate one-year horizon is useful for context but not exactly synchronized with our October 2027 model.
Our $22.75 Base is 21.1% below that dated mean: |22.75 − 28.82| / 28.82. This crosses our 20% review threshold. The gap reflects our retained investment losses, partial fee-yield recovery, normalized tax and larger share allowance, plus a 12× multiple. It does not establish that analysts use the same earnings definition or that their forecasts are wrong. We have not verified a fresh post-release mean and do not reverse-engineer our model to match the old one.
The same action history records TD Cowen maintaining Buy while lowering its target from $26 to $25 on October 6. That is a target cut, not a rating downgrade, and it predates these results. The Fly, via TipRanks, describes a consumer-finance target review ahead of Q3 earnings. We rely on that secondary report, not the original broker research, and do not recast it as a response to the October 29 results.
9. What would change our view?
We would raise the Base after sustained FRLPC recovery supported by realized pricing, investment losses no worse than our allowance, and better cash conversion after retained investments—not merely after a higher volume announcement. We would reduce it if funding tightens, partner growth requires more concessions, losses rise, or capital needs and dilution exceed the model. The next annual filing and quarter-end notes must also confirm financing terms and the changing equity overhang.
Risk is concentrated in the interaction of credit and funding: weaker borrowers can hurt asset marks at the same time that securitization buyers demand more yield. Fair-value estimates depend on assumptions about defaults, prepayments and discount rates. Bank partners can change appetite; regulators and counterparties can demand more compliance, transparency or capital. Technology errors, data misuse, cyber incidents and consumer-treatment failures can damage both distribution and economics. Our scenarios do not exhaust those tail risks.
Concentration disclosures need careful interpretation. The June 10-Q identified two asset-backed securitization (ABS) related-party trusts that together represented 29% of Q2 fee revenue; that is not evidence that two lending banks supplied 29%. The same filing said no single customer exceeded 10% of first-half fee revenue. Vehicle-level revenue concentration, bank-partner concentration and ultimate borrower diversification are different questions.
Related coverage and research record
Our Rocket Companies analysis of mortgage origination and servicing economics explains a related valuation problem: financing assets and debt must be evaluated together, though RKT is not a direct operating comparable to Pagaya. Visit the Company Research archive for other fundamental studies and our research, AI-use and corrections policy for editorial responsibility.
Coverage record: Initial PGY research, October 30, 2026 KST. No earlier AHNIM PGY valuation is overwritten. Future revisions should retain these assumptions and distinguish new information from corrections. This is informational and educational research, not personalized investment advice or a performance guarantee. The cover is an AI-generated editorial illustration, not a photograph of a Pagaya facility.
Sources and original documents
Pagaya — Q3 2026 shareholder letter and unaudited financial statements, released October 29, 2026; quarter and nine months ended September 30, 2026: https://investor.pagaya.com/static-files/038a561b-0a53-45d9-ac6e-920a3406e001
Pagaya — Form 10-Q, quarter ended June 30, 2026; revenue recognition, financing, equity and concentration notes: https://www.sec.gov/Archives/edgar/data/1883085/000188308526000058/pgy-20260630.htm
Pagaya investor relations — Q3 2026 earnings event, October 29, 2026: https://investor.pagaya.com/news-events/events
Pagaya investor relations — SEC filing index, checked October 30 KST; current earnings package distinguished from June 10-Q notes: https://investor.pagaya.com/financial-information/sec-filings
StockAnalysis — PGY market quote, October 29, 2026 regular-session close: https://stockanalysis.com/stocks/pgy/
StockAnalysis / S&P Global — rolling 12-month targets; displayed update October 6, 2026 and last check October 20; retrieved October 30 KST: https://stockanalysis.com/stocks/pgy/forecast/
The Fly via TipRanks — TD Cowen target reduction from $26 to $25 in its Q3 preview; October 6 action cross-checked against the dated forecast history: https://www.tipranks.com/news/the-fly/pagaya-price-target-lowered-to-25-from-26-at-td-cowen-thefly-news
Reviewed October 30, 2026 KST. Company facts and guidance are separated from AHNIM assumptions. Financial statement amounts are converted from thousands to millions where labeled. All model values and analytical tables are recomputed in the accompanying verification record.
Leave a Reply