Upstart Q4 Earnings Call Highlights

Upstart (NASDAQ:UPST) used its fourth-quarter and full-year 2025 earnings call to highlight a year of rapid growth, a return to profitability, and a leadership transition that will take effect May 1. Co-founder Dave Girouard said co-founder and CTO Paul Gu will become chief executive officer, while Girouard will shift to executive chairman.

Leadership transition and executive changes

Girouard said the CEO transition has been planned for years and is aligned with a personal milestone, noting he will remain “quite involved” as executive chairman and continue to help shape strategy and major initiatives. Gu said he is “extremely grateful” for the opportunity to learn from Girouard over 14 years and described Upstart as stronger than at any point in its history.

Gu also outlined several leadership moves tied to the transition. Sanjay Datta will serve as president and chief capital officer, Andrea will join as CFO, and Grant will return to Upstart as CTO.

Full-year 2025: growth, profitability, and balance sheet reduction

Girouard said Upstart grew originations 86% and revenue 64% in 2025 while increasing headcount 18%. He said profitability improved primarily due to market share gains in the personal loan product, where originations grew 75% year over year, and also because newer products expanded rapidly.

According to Datta, Upstart finished 2025 with $950 million in fee revenue and $1.04 billion in total revenue, compared with prior guidance of $920 million and $1.0 billion, respectively. Datta said the company ended the year with positive net income of $54 million and an adjusted EBITDA margin of 22%, exceeding an 18% target discussed a year earlier. Adjusted EBITDA for the year rose to $230 million from $11 million, which Datta attributed to operating leverage and cost discipline alongside investment in new products.

Upstart also emphasized reducing on-balance-sheet exposure even as originations accelerated in newer categories. Datta said loans held on the balance sheet ended Q4 at about $985 million, down from $1.2 billion in Q3, while Girouard said the company reduced balance sheet loans 20% in Q4 and expects that trend to continue.

Q4 results: higher revenue and return to GAAP profitability

Datta said total revenue in Q4 was about $296 million, up 35% year over year and 7% sequentially, including approximately $265 million in fee revenue, up 33% year over year and above quarterly guidance. He attributed the outperformance to the impact of recent underwriting model launches. Net interest income was about $31 million, roughly $5 million ahead of guidance, which Datta said reflected strong returns on an elevated loan balance earlier in the quarter; he added that net interest income should moderate as the company continues to reduce balance sheet exposure.

On the bottom line, Gu said net income swung from a loss of $2.8 million in the prior-year quarter to a profit of $19 million in Q4 2025. Datta reported GAAP EPS of $0.17 on a diluted weighted average share count of 112 million, and adjusted EBITDA of roughly $64 million.

Transaction volume across the platform was approximately 456,000 in Q4, up 86% year over year and 6% sequentially, representing about 307,000 new borrowers. Average loan size was approximately $7,000, up 5% from the prior quarter, which Datta said reflected a higher mix of non-personal-loan products that generally carry larger loan sizes.

Product expansion and funding partnerships

Management repeatedly pointed to momentum in newer secured products. Girouard said auto and home originations both grew 5x year over year in 2025 and accelerated in Q4. Gu added that total loan origination dollars rose 52% year over year in Q4, including 41% growth in personal loan origination dollars, which he described as continued share gains in the 12th year of the personal loan business.

In auto, Gu said Upstart doubled the number of live lending rooftops for the third quarter in a row, launched remote contract signing, and automated document verification with LLMs, reducing funding times by 12%. He said combined auto originations in Q4 were up 56% sequentially and 340% year over year. In home, Gu said Upstart tripled the percentage of HELOCs using automated underwriting and scaled cross-sell campaigns using property data, leading to Q4 home originations up 70% sequentially and 350% year over year. He also said Upstart’s HELOC process is about six days compared with an industry average of 35 days.

Funding diversification was also a focus. Girouard said 70% of auto and home funding in Q4 came from 11 different partners, with an additional 13 signed for the coming year. Datta provided more detail, saying 92% of Q4 auto lending originations across refinance and retail were funded via third parties, and that a majority of Q4 HELOC production was taken by newly signed institutional and lender partners. In Q&A, management said its funding partners span banks, credit unions, and institutional capital, and that secured products are attracting interest from larger banks than those typically focused on unsecured loans.

AI and credit performance; guidance changes and outlook

Upstart leaders emphasized model improvements and credit performance as core differentiators. Gu said Upstart measures itself by “growth, profits, and credit performance,” and described 2025 as progress on all three. He said the average return of the last 12 quarterly vintages exceeded U.S. Treasuries by 608 basis points, with each vintage exceeding Treasuries by at least 270 basis points. Gu also compared the macro backdrop to prior peaks, noting the Upstart Macro Index (UMI) was in the 1.4s for most of the year, versus around 0.8 in 2021, which he said implied statistically identical loans were 43% less likely to default in 2021 than today.

Gu highlighted Q4 technology developments including model updates (Model 24 and Model 25), partner API changes that drove 24% more channel originations quarter over quarter while lowering latency 34%, and a new verification model architecture he said lowers default rates by 0.8%. He also said training data surpassed 100 million borrower repayment events for the first time.

Upstart also announced changes to how it communicates performance and forecasts. Girouard said the company will begin publishing monthly transaction volume on its platform at the start of each subsequent month and will focus on annual guidance updated as it reports each quarter.

For 2026, Datta said guidance assumes a “constant default risk environment” with UMI holding around 1.4–1.5 and a static interest rate environment. Under that scenario, Upstart expects about $1.4 billion in total revenue, approximately $1.3 billion in fee revenue, and an adjusted EBITDA margin around 21%. Datta also said auto and home are expected to contribute over $100 million of fee revenue in 2026, and outlined expectations for secured products at scale, including average upfront take rates around 4% plus average servicing rates around 2% of outstanding balance on average loan sizes of roughly $30,000.

Management said contribution margins have declined intentionally due to mix shift toward secured products and prime borrowers with lower take rates and due to increased emphasis on customer lifetime value, which they said can lower current take rates while supporting higher volumes and long-term profits. Datta said Upstart aims for absolute contribution dollars to grow robustly and targeted keeping contribution dollar growth within at least five percentage points of fee revenue growth.

Looking longer term, Girouard said Upstart projects a 35% compound annual growth rate for the next three years. Datta added that for 2025 to 2028, the company is targeting a 35% CAGR in a “macro-neutral” environment and a terminal adjusted EBITDA margin of about 25% in 2028, which management said would be driven largely by operating leverage and internal productivity gains from AI.

About Upstart (NASDAQ:UPST)

Upstart Holdings, Inc operates a cloud-based lending marketplace that leverages artificial intelligence and machine learning to assess borrower creditworthiness. The company partners with banks and credit unions, providing its proprietary AI models and underwriting platform to facilitate consumer credit products. By focusing on non‐traditional data points—such as education, employment history and other real‐time indicators—Upstart seeks to improve approval rates and lower loss rates compared with conventional credit scoring methods.

Upstart’s core offering centers on unsecured personal loans, which borrowers can use for purposes such as debt consolidation, home improvements or major purchases.