Upstart's Q3 2025: Contradictions Emerge on Model Conservatism, Macroeconomic Strategy, Auto Lending, and Funding

Date of Call: November 4, 2025

Financials Results

• Revenue: $277M total revenue for Q3, up 71% YOY and 8% sequentially (fees ~$259M, up 54% YOY; net interest income ~$19M)
• EPS: $0.23 GAAP diluted EPS (based on 110M diluted shares); adjusted EPS $0.52 (based on 125M diluted shares)

Guidance:

• Q4 2025: total revenue ~ $288M (fees ~$262M; net interest income ~$26M).
• Q4 contribution margin ~ 53%; GAAP net income ~ $17M; adjusted net income ~ $52M; adjusted EBITDA ~ $63M (diluted ~111M shares).
• FY2025: total revenue ~ $1.035B (fees ~$946M; NII ~$89M); adjusted EBITDA margin ~ 22%; GAAP net income ~ $50M.

Business Commentary:

* Increased Transaction Volume: - Upstart reported a 128% increase in loan transactions across their platform year-on-year, with approximately 428,000 loans in Q3. - The growth was driven by strong consumer demand, with more than 2 million applications submitted, up 30% from Q2, despite a model taking a more conservative approach in Q3.

• Model Conservatism and Conversion Rates:
• Upstart experienced a reduction in the conversion rate from 23.9% in Q2 to 20.6% in Q3, due to the model's response to macroeconomic signals, such as a modest rise in the Upstart Macro Index (UMI).

• The model's increased caution led to fewer approvals and higher interest rates, impacting transaction volume.

• Auto and Home Business Expansion:

• The auto retail business grew significantly, with the number of live lending rooftops doubling in Q3 compared to the previous quarter.

• Transaction volume for auto retail increased by more than 70% sequentially, driven by expansion into new states and improvements in software.

• Funding and Partnerships:

• Upstart has a strong position in core business funding, with significant excess capacity, and achieved a new monthly high in available funding from partners in Q3.
• The company is focusing on reducing R&D-related balance sheet holdings and intends to shift to third-party funding for newer products.

  Sentiment Analysis:

  Overall Tone: Positive

  • Management highlighted 80% YOY transaction volume growth (2M applications), Q3 GAAP net income of ~$32M, 71% revenue growth, and repeated optimism about AI leadership and an "amazing 2026," emphasizing strong funding partnerships and continued product momentum.

Q&A:

• Question from Dan Dolev (Mizuho Securities USA LLC): Can you comment on the very strong application demand in Q3 and reconcile that with guidance being below expectations?
  Response: Applications rose ~30% Q/Q, but model conservatism (responding to short-term macro signals) reduced approvals/conversion in Q3, which constrained volume; models have since reverted.

• Question from Kyle Peterson (Needham & Company, LLC): Have recent negative auto headlines impacted your auto expansion plans or customer conversations?
  Response: No material impact; Upstart hasn't seen widespread fraud, underwriting processes mitigate dealer risk, though partner diligence has increased modestly.

• Question from Kyle Peterson (Needham & Company, LLC): Was sequential decline concentrated in super-prime (720+) versus the core product—i.e., was model tightness focused in that segment or is competition a factor?
  Response: It's a mix: model tightening affected primary segments including low-to-mid 700s (higher UMI there), and heightened competition/pricing also weighed on that segment.

• Question from Peter Christiansen (Citigroup Inc.): With the marketing/channel improvements and more applications, what's the quality of these leads—same, improved, or worse?
  Response: Marketing generated more applications, but conversion fell mainly due to temporary model conservatism, so the higher application counts mechanically included more applicants less likely to convert under the tightened model.

• Question from Peter Christiansen (Citigroup Inc.): If non-prime auto delinquencies improve, would that meaningfully move Upstart's auto originations?
  Response: Yes—improvements would be a tailwind; Upstart sees good auto credit performance and views market disruption as opportunity to scale auto in 2026.

• Question from Simon Alistair Clinch (Rothschild & Co Redburn): Are you assuming model conservatism continues into Q4 despite UMI starting to improve?
  Response: Yes—UMI improvements are materializing but management remains conservative and expects some Q4 impact from prior conservatism due to lagging model response.

• Question from Simon Alistair Clinch (Rothschild & Co Redburn): Is personal loan demand still primarily refinancing credit card debt or broadening to other uses?
  Response: Refinancing remains the dominant use case, but unsecured personal loans are broadly used and increasingly compete with secured options for many purposes.

• Question from Patrick Moley (Piper Sandler & Co.): How have funding-partner conversations trended for R&D products and has private credit appetite contracted given auto headlines?
  Response: Conversations are progressing well with healthy appetite; deals are large and diligence-heavy so timelines vary and have lengthened, but no apparent contraction in demand.

• Question from Mihir Bhatia (BofA Securities): Was higher UMI the primary driver of the conversion decline, and how will you limit conversion volatility going forward?
  Response: Yes—the dominant driver was model conservatism reducing approvals, rates, and loan sizes; calibration improvements implemented this quarter are expected to cut macro-driven month-to-month conversion volatility by ~50%.

• Question from Mihir Bhatia (BofA Securities): Repayment speeds have increased—what's driving that and credit implications/delinquency impact?
  Response: Repayment speeds rose broadly (partial and full prepayments), likely reflecting improving consumer health; this reduces near-term interest income and led models to price conservatively, but is credit-positive longer term.

• Question from Reginald Smith (JPMorgan Chase & Co): Was the conversion decline driven more by approvals or consumer acceptance, and do you see share shifts versus competitors?
  Response: Predominantly approval-side driven (fewer approvals); management sees no definitive evidence on competitor mix/strategy differences—Upstart follows model signals rather than overriding them.

• Question from Reginald Smith (JPMorgan Chase & Co): Did declines occur in super-prime?
  Response: No—super-prime saw rate increases rather than declines; most declines occurred among lower-score borrowers.

• Question from Reginald Smith (JPMorgan Chase & Co): What day-1 economics/take rate should we expect for HELOCs versus core PL?
  Response: HELOC take rates expected to be healthy but meaningfully lower than PL (roughly ~half), while average loan sizes are substantially larger (more than double).

• Question from Reginald Smith (JPMorgan Chase & Co): Any credit performance issues in your portfolio given the UMI activity?
  Response: No—credit performance remains strong and the UMI is used to make disciplined, prudent adjustments on behalf of lenders.

• Question from James Faucette (Morgan Stanley): Which model elements weakened and subsequently improved; where was the model looking?
  Response: Model separation remains best ever; the calibration side became conservatively responsive to short-term macro signals (sampling/measurement noise), and recent fixes reduced that noise materially.

• Question from James Faucette (Morgan Stanley): How should we think about exit rates into Q4 and run rates into 2026?
  Response: Management is optimistic—expects conversion recovery by quarter end as model and targeting improvements roll out and views 2026 as a strong growth year.

• Question from Robert Wildhack (Autonomous Research US LP): Why does UMI show subprime in better shape while mid-prime looks worse vs broader headlines?
  Response: UMI varies by score band: sub-660 (lower end) shows modest UMI, mid-prime (low–mid 700s) exhibits higher UMI vs pre-COVID, and 800+ cohorts are doing well—resulting in a U-shaped pattern across score bands.

• Question from Robert Wildhack (Autonomous Research US LP): What drove sequential declines in Engineering and G&A OpEx?
  Response: Fixed-cost discipline and mechanical reductions (lower bonus/comp accruals tied to reduced outlook) drove sequential declines in those OpEx lines.

• Question from John Hecht (Jefferies LLC): HELOC average size ~$55k–$60k—what are borrower use cases?
  Response: HELOCs are general-purpose (commonly home improvement or debt repayment); company provided no detailed use-case breakout.

Contradiction Point 1

Model Conservatism and Application Volume Growth

It involves differing explanations of the model's conservatism and its impact on application volume growth, which are crucial for understanding the company's growth strategy and risk management.

Can you explain the strong demand in Q3 and link it to guidance? - Dan Dolev(Mizuho Securities USA LLC, Research Division)

2025Q3: The application volume grew about 30% quarter-on-quarter, which was ahead of the transaction volume. The model took a step towards conservatism seeing macro factors, a natural thing, and even overreacted. - David Girouard(CEO)

How has Upstart responded to macroeconomic uncertainty? - Mihir Bhatia(Bank of America)

2025Q1: We're conservative in our planning and rely on models to adapt to macro changes. We build in conservatism to account for potential disruptions. - David Girouard(CEO)

Contradiction Point 2

Macroeconomic Impact and Business Strategy

It involves differing perspectives on the impact of macroeconomic factors on the company's business strategy, which could influence investor expectations and growth prospects.

With the increase in UMI, will the model's conservatism persist into Q4? - Simon Alistair Clinch(Rothschild & Co Redburn, Research Division)

2025Q3: The recent improvements in UMIs are materializing, but there's some lag. They will impact Q4. We are cautious and want to ensure conservatism in our model. - Sanjay Datta(CFO)

What factors are influencing contribution margins, and how will they evolve throughout the year? - Simon Clinch(Redburn Atlantic)

2025Q1: We track macroeconomic factors, some of which are macroeconomic factors and some of which are economic factors that we think are relevant to our business. And of course, as you can see, we've been running a model to reflect the macroeconomic factors. - Sanjay Datta(CFO)

Contradiction Point 3

Auto Lending Business and Partnerships

It involves the potential impact of recent credit issues in the auto space on the company's expansion plans and partnerships, which can affect business growth and strategic direction.

Have recent bankruptcies affected your expansion plans or customer conversations in the auto industry? - Kyle Peterson(Needham & Company)

2025Q3: None of the recent issues affected us. We've handled these incidents by building strong underwriting processes to ensure we're effectively mitigating risks against dealer activity. - David Girouard(CEO)

How should we assess your operating leverage for the medium- and long-term? - Dan Dolev(Mizuho)

2024Q4: Conversations are progressing well with multiple agreements expected soon. The auto space has increased diligence due to recent issues, but our partners are excited about our platform. - Sanjay Datta(CFO)

Contradiction Point 4

Funding Availability and Capital Strategy

It involves the company's approach to securing funding and capital, which is crucial for supporting business growth and maintaining financial stability.

How is progress with R&D funding partners, and any impact from auto industry credit issues on them? - Kyle Peterson(Needham & Company)

2025Q3: We've made significant progress deploying capital against the agreement, and we'll have a nice tailwind from these agreements as we move into next year. - Sanjay Datta(CFO)

How to approach the funding mix between committed capital and at-will buyers going forward? - Kyle Peterson(Needham)

2024Q4: We're in conversations to create capital sources to take loans from our balance sheet. It's more of a quarter-to-quarter activity than a monthly one. - Sanjay Datta(CFO)

Contradiction Point 5

Model Conservatism and Performance

It involves changes in the company's risk management and model performance, which can impact the financial outlook and investor confidence.

Can you explain the strong Q3 demand and how it relates to the guidance? - Dan Dolev(Mizuho Securities USA LLC)

2025Q3: The model took a step towards conservatism seeing macro factors, a natural thing, and even overreacted. - David Girouard(CEO)

What drove the increase in borrower demand this quarter? - Simon Clinch(Rothschild & Co Redburn)

2024Q4: Approximately half comes from model improvements... Our technology and model innovations have enhanced risk separation and platform pricing. - Sanjay Datta(CFO)