Philip A. Fisher Framework Date: 2026-02-22 Atlas Baseline: Conviction 2/5, Gate FAIL. Read and considered.
Before touching the financials, I must report what the mosaic of external voices reveals. This is the foundation.
Credit model critics (most important signal): Iceberg Research published a short report in February 2025 titled "Pagaya: Using Other People's Money to Hide Massive Losses." The allegation is that Pagaya's ABS structure — specifically the almost fully horizontal first-loss retention — obscures cumulative credit losses by spreading them across marked-to-market "fair value" estimates that management controls. A separate, credible industry publication (Fintech Takes, Feb 21 2025) published "Pagaya Doesn't Make Sense To Me," questioning the fundamental partner economics. These are not casual dismissals; they represent informed industry practitioners raising structural questions about the business model. I cannot fetch the full texts, but the headlines and search snippet summaries are revealing in themselves. Companies that generate substantive skeptical analysis from credible credit practitioners warrant extra scrutiny on Points 10 and 14.
Data from 29 publicly rated ABS transactions shows multiple lowest-quality tranches have been downgraded or are unrated, with Pagaya reportedly buying back collateral to prevent further downgrades. This is precisely the kind of behavior that does not appear in income statements.
Employee intelligence: Pagaya laid off 20% of its 600-person workforce in January 2025. This was the second round in eighteen months — the first being 140 of 650 employees in early 2023. Those close to the company framed both rounds as "accelerating the transition to profitability." The Glassdoor rating is 3.2/5. Two layoff rounds in eighteen months at a company ostensibly in a growth phase is not an efficiency story; it is an organization that has been repeatedly miscalibrated on headcount relative to revenue expectations. This is a direct signal on Point 7 (labor relations) and Point 9 (management depth).
Competitor landscape (PGY vs UPST): Pagaya and Upstart are the two principal AI-driven lending platforms. The structural difference is meaningful: Upstart operates closer to a marketplace model (more credit cycle exposed, more volatile); Pagaya is B2B2C, embedded in partner lender workflows, capital-lighter on origination. Pagaya's ROE of 44% exceeds Upstart's 24%. However, Upstart guided FY26 revenues of $1.4B with 62.8% FY25 growth versus Pagaya's 19.2% FY26 consensus. Upstart's faster growth at comparable revenue scale undermines the claim that Pagaya's model is superior on growth generation. The more defensible claim is that Pagaya is structurally stickier with partners. Whether stickiness translates to growth is the unanswered question.
Analyst consensus: Wall Street is broadly bullish: 8 Buy, 1 Hold, 0 Sell. Citi upgraded to Buy. Benchmark has a $25 target. Zacks consensus implies $2.65 EPS FY25 and $3.43 FY26 (29% growth). These are optimistic relative to management's own 15% revenue guidance. The consensus enthusiasm from sell-side analysts — who are not structurally incentivized to ask the hard credit questions — should be discounted heavily in light of the practitioner-level skepticism noted above.
CEO assessment: Gal Krubiner is an Israeli co-founder who built Pagaya with two partners — one with hard-core data science credentials, one with capital-raising expertise. The founding team is intact, which is a positive signal for alignment. McKinsey published a long-form interview with Krubiner articulating the vision clearly. He is evidently thoughtful and media-savvy. However, intellectual clarity about a vision is distinct from operational execution — and the transparency failure around the 2024 credit impairments is the relevant data point, not the McKinsey profile.
Hiring patterns: Despite the January 2025 layoffs, ZipRecruiter shows 10 active Pagaya positions as of October 2025, indicating targeted rebuilding in specific areas. This pattern — mass layoff followed by selective rehiring — is consistent with a company that over-hired in a growth phase, cut to profitability, and is now building selectively. It is not consistent with an organization that has a stable, well-managed human capital structure.
I will work through the Points that are most material, then summarize the remainder.
Point 1 — Products/services for substantial sales growth over several years: The product is an AI-driven credit underwriting engine, embedded in partner lenders' workflows. The theoretical runway is vast: the US consumer lending market is enormous, and if the AI model genuinely identifies creditworthy borrowers that traditional FICO-based scoring rejects, the value proposition is durable. Three new partners onboarded in Q4 (Achieve, Global Lending Services, a major unnamed BNPL provider). These are not tier-1 bank names; they are mid-tier fintech lenders. The partner base expansion is occurring, but at a pace that generated 20% YoY growth — not the 40%+ one expects from a company with a genuinely dominant product in an early-capture phase. Rating: Acceptable but decelerating.
Point 2 — Management determination to develop new products: The revolving ABS structure launched in FY25 is a genuine product innovation — it replaces episodic term ABS deals with a continuous funding facility, improving capital efficiency and reducing Pagaya's balance sheet exposure over time. This is management thinking ahead about their capital structure, not just their revenue line. It was oversubscribed 30% above initial sizing, which is a positive market signal on the ABS investor side. However, the core product — the AI underwriting model — has not demonstrably evolved. I have found no evidence of meaningful new product lines, new geographies, or new verticals. Rating: Modest.
Point 3 — Research and development effectiveness: This is the most important and most opaque point for Pagaya. The AI model is the product. By definition, ongoing R&D should improve model accuracy, reduce loss rates on retained tranches, and expand the addressable population of creditworthy-but-mispriced borrowers. The 2021-2023 vintage impairments ($229M charge in Q4 FY24) are a direct test of R&D effectiveness — and the model failed to anticipate the deterioration in those cohorts. Management's explanation is that COVID-era distortions, stimulus effects, and post-COVID normalization created an unprecedented environment that any credit model would have struggled with. This is partially plausible — Upstart had the same problem. But the scale of the impairment ($229M on a company with $81M total FY25 GAAP income) suggests the model had a fundamental mispricing of credit risk, not merely a one-standard-deviation prediction error. Separately, after two layoff rounds, one must ask: what happened to the research organization? Two rounds of 20% reductions in eighteen months have almost certainly reduced the team that builds the model. Rating: Uncertain — the evidence cuts both ways, but the impairment scale is a yellow flag.
Point 4 — Above-average sales organization: Pagaya sells B2B — to lenders, not borrowers. The partner relationships appear sticky (switching costs are high once Pagaya's underwriting is integrated into a lender's origination system). Three new partners in Q4 is a positive signal. The ABS investor base (oversubscribed 30%) is also a customer of sorts — capital markets counterparties that must be cultivated and retained. The explicit focus on diversifying the ABS investor base in 2025 shows management's awareness of this dependency. Rating: Acceptable.
Point 5 — Profit margin trajectory: Adj EBITDA margin improved from ~17% in FY24 to ~29% in FY25. GAAP net income turned positive across all four quarters of FY25 ($81M for the year). This is a genuine improvement. However, the gross margin proxy (FRLPC basis, ~39%) is structurally thin for a company that bills itself as an AI software platform. The comparison to Upstart (which has similar ABS exposure and similarly thin GAAP margins) is instructive — these are not SaaS businesses with 70%+ gross margins; they are AI-enhanced financial intermediaries with credit risk embedded in their economics. The margin improvement from 17% to 29% is impressive, but the absolute level is consistent with a financial services company, not a technology company. Rating: Improving from a low base.
Point 6 — Maintaining profit margins: The Q4 deliberate pullback from high-risk credit tiers (~$100-150M of volume foregone) is the key management decision this quarter. CEO Krubiner frames it as "becoming a better Pagaya, not just a bigger one." I have seen this exact framing from management teams before. Sometimes it is genuine strategic discipline; more often it is post-hoc rationalization of credit deterioration that forced the pullback. The honest question is: did management choose to pull back, or did ABS investor demand for high-risk tranches dry up, effectively making the pullback involuntary? The ABS oversubscription at 30% above initial sizing suggests capital markets appetite is strong — which means the pullback was more likely management choice than market constraint. If so, this is a genuine discipline signal. I am giving management marginal benefit of the doubt here, but I am watching. Rating: Ambiguous.
Point 7 — Labor and personnel relations: Two 20% layoffs in eighteen months. Glassdoor at 3.2/5. I have found this pattern in companies I have studied over many decades: when an organization repeatedly surprises itself on headcount requirements, it either lacks planning discipline or is operating in a business environment so volatile that planning is impossible. Neither interpretation is compatible with a "hold almost never" type of business. Fisher's ideal company — a Motorola or a Texas Instruments or a Dow Chemical in its best years — maintained its people through cycles because the people were the asset. Pagaya is treating its workforce as a variable cost, not an asset. Rating: Weak.
Point 8 — Executive relations: The founding trio (Krubiner, Pardo, Yulzari) is intact. President Sanjiv Das (former CitiMortgage CEO) adds meaningful consumer finance credibility. CFO Evangelos Perros has been consistent on calls. The executive team is experienced and appears aligned. Rating: Acceptable.
Point 9 — Management depth: After two rounds of 20% layoffs, this is structurally impaired. I cannot assess from the outside how many senior individual contributors were lost, but historical base rates suggest layoffs of this magnitude destroy institutional knowledge disproportionately. The experienced "middle" of an organization — the people who know where the bodies are buried — is typically hit hardest. Rating: Degraded.
Point 10 — Cost analysis and accounting controls: This is a serious concern. The $229M fair value impairment on risk-retention notes in Q4 FY24 was not adequately telegraphed to investors in preceding quarters. Multiple sources confirm that management's narrative about credit problems being confined to early vintages was contradicted by the scale and timing of the write-down. Fair value marks on retained ABS tranches are management estimates — not arm's-length transactions. The potential for these marks to lag actual credit deterioration is structural, not coincidental. Rating: Weak.
Point 11 — Industry competitive position: Pagaya's Dodd-Frank-mandated risk retention (above the 5% minimum, fully horizontal structure) creates a structural bind: it forces Pagaya to retain first-loss tranches, which is simultaneously their competitive moat (partners value the risk transfer) and their greatest vulnerability (losses are amplified in downturns). This is not a software business — it is a credit intermediary wearing software clothing. The #1 personal loan ABS issuer position is a genuine competitive strength, but it is not the moat that compounding technology advantages create. Rating: Structurally complex.
Point 12 — Long-range profit outlook: FY26 guided at 15% revenue growth midpoint (1.49Bmidpoint).Thetrajectory : FY24 1.03B (+26%), FY25 ~1.30B(+261.49B (+15%). This is a decelerating growth arc. Three new partners onboarded in Q4, but they ramp slowly — partner economics take 12-24 months to fully activate. The credit impairment reserve of $125M guided for FY25 acts as a recurring drag on GAAP earnings even as the operating business improves. Rating: Decelerating.
Point 14 — Management frankness with investors during adversity: The Q4 FY24 impairment episode is the controlling data point. Management did not adequately warn investors before the charge crystallized. On this quarter's call, the CEO's language ("becoming a better Pagaya, not just a bigger one") is elegant but evasive — it does not directly address whether the credit model has been structurally improved or whether the pullback is permanent or temporary. I prefer management that says: "The model mispriced these credit cohorts because X. Here is what we changed. Here is the test we will apply to know if the change worked." That is not what I am reading in the transcript summary. Rating: Weak.
Point 15 — Unquestionable management integrity: No proven fraud. The Iceberg Research short report makes allegations, but short sellers have incentives to exaggerate. I will not convict management on a short report alone. However, the transparency failure on the FY24 impairment — repeated assurances of "disciplined capital allocation" while investing $566M in notes that subsequently required $229M of write-downs — creates a reasonable question about whether management's public communications lag their internal awareness of credit deterioration. This is not integrity in the sense of outright dishonesty; it is integrity in the more subtle sense Fisher meant: do they tell shareholders what they know, when they know it? Rating: In question — not disqualifying, but cannot award a clear pass.
Is Pagaya "fortunate because it is able" or "fortunate because it is fortunate"?
The honest answer is: primarily the latter, with elements of the former.
Pagaya grew 26% in both FY24 and FY25, riding the consumer credit expansion of the post-COVID normalization period and the personal loan market's recovery. When credit conditions tightened and delinquency risk rose in certain borrower segments, Pagaya had to reduce volume — not because it chose to at the top of a growth arc, but because those credit segments became uneconomical to serve at the yields available in ABS markets. The pullback is market-imposed, not strategy-driven, regardless of how management frames it.
The transition to a "better Pagaya" framing could represent genuine strategy evolution — focusing on higher-quality partners, better-credit borrowers, more sophisticated ABS structures — but that transition guides revenue growth down to 15%. A truly "fortunate because able" company would use its proprietary AI advantage to expand into new credit segments as old ones tighten, not to contract.
The three new partners added in Q4 are encouraging as proof-of-product signs, but they are not the volumes that a dominant-moat AI credit underwriter should be generating.
Prior beliefs (no prior coverage — establishing baseline): No prior Phil Fisher analysis on PGY. Starting fresh.
Updated beliefs after this review:
Avoid.
Not on valuation grounds — at ~1x TTM revenue and ~8x forward earnings, Pagaya is objectively cheap. Atlas is correct that the market is pricing in a worse outcome than the base case warrants.
But Fisher did not invest in cheap companies. He invested in outstanding companies at fair prices. Pagaya does not qualify as an outstanding company. The Fifteen Points are the standard, and Pagaya fails on: labor relations (Point 7), management depth (Point 9), accounting transparency (Point 10), long-range profit outlook direction (Point 12), and management frankness (Point 14). Point 15 carries an asterisk.
More fundamentally, the growth arc is moving in the wrong direction. My forty years of experience have taught me that companies with decelerating growth — from 36% to 20% to a guided 15% — in a market that is still expanding are losing competitive position, not gaining it. The "deliberate pullback" framing may be accurate, but a company with a truly superior AI credit model would use the pullback in high-risk segments as an opportunity to expand in high-quality segments. The net result would be a growth rate near or above the prior trend. At 15%, that is not what is happening.
The ABS impairment history and the transparency failures around it disqualify Pagaya from the "hold almost never" category regardless of current price. When management does not tell shareholders what it knows as soon as it knows it — the most important criterion of all — I must set aside the valuation argument and pass.
I will track this company. If the FY24-vintage impairments resolve without further write-downs, if Q1/Q2 FY26 partner volume shows material acceleration from the three new Q4 partners, and if management produces two consecutive quarters of unambiguous transparency on credit performance, the case for reconsideration becomes meaningful. Until then: Avoid.
Analysis saved: Q4 FY25 | Date: 2026-02-22 | Status: New coverage — Avoid