Upstart Holdings, Inc. (NASDAQ: UPST)
Prepared Remarks: Q1 2025 Earnings Call
Tuesday, May 6, 2025, 1:30pm PST
For more, please see Upstart’s Q1 2025 earnings press release and presentation.
Participants:
Dave Girouard, Co-founder and CEO
Sanjay Datta, CFO
Ross Monro, SVP Finance
Sonya Banerjee, Head of IR
Operator:
Good afternoon and welcome to the Upstart first quarter 2025 earnings call. At this time, all participants are in listen-only mode to prevent any background noise. Later, we will conduct a question-and-answer session, and instructions will be given at that time. As a reminder, this conference call is being recorded. I would now like to turn the call over to Sonya Banerjee, Head of Investor Relations. Sonya, please go ahead.
Sonya Banerjee, Head of IR
Thank you. Welcome to the Upstart earnings call for the first quarter of 2025. With me on today’s call are: Dave Girouard, our co-founder and CEO, and Sanjay Datta, our CFO.
During today’s call we will make forward-looking statements, which include statements about our outlook and business strategy. These statements are based on our expectations and beliefs as of today, which are subject to a variety of risks, uncertainties, and assumptions, and should not be viewed as a guarantee of future performance. Actual results may differ materially as a result of various risk factors that have been described in our SEC filings. We assume no obligation to update any forward-looking statements as the result of new information or future events, except as required by law.
Our discussion will include non-GAAP financial measures, which are not a substitute for our GAAP results. Reconciliations of our historical GAAP to non-GAAP results can be found in our earnings materials, which are available on our IR website.
And with that, Dave, over to you.
Dave Girouard, Co-founder and CEO
Thanks, Sonya. Good afternoon, everyone. Thank you for joining us today.
The first quarter was a strong one for Upstart, despite being our seasonally slowest time of the year. Platform originations grew 89% year-on-year, with model wins and improved borrower health combining with more competitive capital to drive meaningfully higher conversion rates.
Our revenue grew 67% year-on-year and our Adjusted EBITDA margin reached 20% for the first time in three years. With fees generating 87% of our revenue, we were also right on the doorstep of GAAP profitability in Q1. Just as importantly, Home and Auto continued their torrid growth pace, with originations growing 52% and 42%, respectively, on a sequential basis.
From our perspective, consumer financial health, as represented by the Upstart Macro Index, has been largely improving for almost a year now. Currently it remains elevated but stable. Our credit continues to perform well and, while we’re vigilant with respect to any disruptions that recent government trade policy might cause, we’re also confident that our ability to adapt to changing macroeconomic conditions is miles better than it was just a couple of years ago. In a time of trade disruption, we’re also happy that Upstart is a 100% U.S. business that’s 100% digital.
Now I’ll take you through some of the progress we made in each of our product areas in Q1.
Product progress
With our core personal loan product, originations were flat sequentially and up 83% year-over-year. We continue to make rapid improvements to our models that facilitate underwriting and automated approvals. Additionally, we continue to strengthen and diversify the capital supply that is the fuel of our business. These improvements have contributed to increasing our conversion rates from 14% a year ago to 19% in Q1. We also reached an all-time high of 92% of loans fully automated – meaning the entire process from rate request to loan closing is entirely driven by AI-powered software, with no human intervention by Upstart. All else being equal, we believe a faster automated process selects for better borrowers.
Additionally, our “best rates, best process for all” mantra has really begun to pay dividends. In Q1, 32% of our originations were to super prime borrowers, which we – as well as the CFPB – define as borrowers with credit scores higher than 720. We measure progress against “best rates” not just by portfolio mix but also by win rates versus competitors across the credit spectrum. In this vein, we’ve also made extraordinary strides in recent months. I’m excited to share more about this with you, which I expect to do at our AI Day event next week.
We continued to stack up model wins, putting more distance between Upstart AI and the rest of the industry. Since last we spoke, we introduced embeddings to our core personal loan underwriting model. Embeddings are a machine-learning technique that convert complex, unstructured data into useful model inputs, or “features” in ML-speak. This is done by clustering data that have meaningful relationships, allowing seemingly random data to become valuable to predicting credit performance.
For example, two consumers may have different credit cards – say, an Amex and a Chase Sapphire card. But with embeddings, our model can learn that these cards might reflect similar consumer behaviors. What makes this approach so powerful is that embeddings help us uncover subtle patterns that would be difficult – or even impossible – to identify otherwise. This leads to better model generalization, improved accuracy and more informed credit decisions. While embeddings are widely used in other domains – like natural language processing – applying them to credit underwriting is entirely novel. We’re excited about what embeddings can do to drive our risk separation metrics forward, and we’re equally excited about the pipeline of modeling innovations in front of us in 2025.
Auto
Last quarter I told you that our auto business finished 2024 by growing originations about 60% sequentially. Well, in the first quarter of 2025 originations grew another 42% sequentially, despite Q1 being the seasonally softest time of the year. And Upstart auto lending grew almost 5X compared to a year ago. These increases were driven principally by model updates and pricing improvements. At the same time, our continued focus on cross-selling existing customers reduced acquisition costs for our auto refinance product by 57% quarter-over-quarter. This mega improvement to CAC was driven – as usual – by improved conversion rate, which more than doubled sequentially in Q1.
In Q1, we also saw our first instant approval of an auto refinance loan, where the borrower completed the entire process in a single session of just 9 minutes. As far as we know, this could be the first instantly approved auto refinance loan in the world. This is a modest beginning but sets us down the path of automation that has been so central to our success in personal loans.
HELOC
In the Home lending category, we’re thrilled with how quickly our HELOC product is maturing. In December, we ported our personal loan models for instantly verifying income and identity to our HELOC product. This increased instant income and identity verification from less than half of loans in Q4 to nearly two-thirds of HELOCs this past quarter. This is an experience borrowers love: Instantly approved applications convert at more than twice the rate of those requiring manual intervention. It’s also a strong demonstration of how our core technology can be generalized across credit categories.
In Q1, we also finally launched the Upstart HELOC in California, bringing our footprint to 37 states plus Washington, D.C., now covering almost 75 percent of the U.S. population.
As I mentioned earlier, in Q1 our HELOC originations grew 52% quarter on quarter and more than 6X compared to a year ago. We now have agreements signed with three lending partners for our HELOC product and have begun the process of moving funding off of our balance sheet. We expect this will take considerable time as we bring both depository and institutional capital to this category.
SDL
Our small-dollar product continues to perform well, with originations growing 7% sequentially and almost tripling year-on-year. Our SDL continues to be a critical customer acquisition tool, accounting for nearly 16 percent of new borrowers on Upstart in Q1.
As I previewed back in February, in Q1 we moved to a single underwriting model for both of our unsecured products. This means the small-dollar product is benefiting straight away from machine-learning innovations such as embeddings that I described earlier.
Servicing
In Q1 we continued work to modernize and scale our servicing operations. We’re rapidly automating routine tasks like processing payment failures and check handling so we can spend human time on more valuable tasks. In Q1, we automated 90% of hardship applications, making the process more seamless for borrowers and more efficient for Upstart.
Beyond the technology, the work we’ve done to prioritize direct collections efforts for borrowers at risk of default have continued to have a meaningful impact. For example, in Q1 we realized a 50% increase in debt settlement acceptances by extending repayment terms for at-risk borrowers. In our auto business, we doubled our recovery rates year-over-year in Q1.
We’ve been improving our data collection and structuring in servicing for quite some time now, and expect to launch our first machine-learning model in this area very soon. We’re excited about the potential for ML in loan servicing to increase efficiency and reduce loss rates. This is a necessary step toward launching our servicing and collections as a highly differentiated standalone offering in the market, which we hope to do in the future.
Progress against 2025 priorities
I’d like to quickly touch on Upstart’s 2025 priorities that I mentioned to you in February. We’re making great strides against this list, and I’m grateful for the many Upstarters who are putting their all into making this an incredible year for Upstart.
First, 10X our leadership in AI. Already this year we’ve made great leaps forward both in process and substance that reinforce that, when it comes to AI lending, Upstart is a category of one. Embeddings are a real breakthrough for our AI foundation model, and I’m convinced it will pay dividends across all of our products over time. Our pipeline of modeling wins is beyond robust, so I couldn’t be happier about our progress here.
Second, prepare our funding supply for rapid growth. We continued to have well over 50% of our loan funding in committed partnership agreements. In early April we signed our first committed capital arrangement with Fortress who is an industry leader in private credit. Looking forward, we expect to add our newer products to these partnerships which will greatly expand on the value and efficiency they bring to Upstart. Additionally, we added 15 new lending partners to our super prime offering, growing capital 38 percent quarter to quarter for that market segment.
Third, return to GAAP net income profitability in the second half of the year. We’re on track for this right now and excited to demonstrate the leverage in our business as both our core and newer products expand rapidly through the year.
And last but not least, giant leaps toward best rates, best process for all. Earlier this year, I challenged each of our product GMs to have the best rates against major market competitors for each borrower segment they serve by the end of 2025. Specifically, this means that our win rate should be higher than any other digital competitor in each of our products. This of course comes with the prerequisite of on-target or better credit performance as well as solid unit economics. I’m pleased to share that we’re running well ahead of this target, particularly in our core personal loan product. You’ll hear more about this next week at AI Day.
To wrap up my remarks today, I think much of the world has come down with a case of AI fatigue. There’s just so much discussion and so many predictions about what AI will do – or what it might do – to the world, to our lives, and to our children’s lives. I admit I roll my eyes at some of the debates and the discussions as well.
But in Upstart’s case, AI and its unique capabilities are unquestionably aligned with a better future for all when it comes to financial wellness. Improvements to our risk models and expansion of our product line mean we are reducing the price of credit for the world more and more each day. If you’re still unclear about the incredible opportunity for AI-enabled lending and Upstart’s leadership in this vital category, I hope you’ll join us next week for AI Day. I promise you won’t be disappointed.
Thank you, and I’d now like to turn it over to Sanjay, our chief financial officer, to walk through our Q1 2025 financial results and guidance. Sanjay?
Sanjay Datta, CFO
Thanks, Dave. And thanks to all of our participants for sharing some of your time with us today.
Financially, the first quarter of 2025 came in slightly ahead of expectation, with outperformance on the top and bottom line owing primarily to higher than anticipated net interest income.
The impact of the typically soft tax seasonality in Q1 played out largely as expected, and despite recent turbulence in the financial markets, the effects of the macroenvironment on credit performance has in our estimation remained relatively stable throughout the quarter and roughly consistent with our prior assumptions.
Total revenue for Q1 came in at approximately $213 million, up 67% year on year.
This overall number included revenue from fees of $185 million, which was up 34% year over year. This amount was in line with guidance, although from a mix perspective we outperformed our expectations in the Super Prime borrower segment, resulting in higher overall origination volumes than anticipated at lower take rates.
Additionally, net interest income represented roughly $28 million of overall revenue, exceeding our outlook by $13 million. Approximately half of this amount came in the form of higher net interest spreads from our balance sheet as we passed peak chargeoffs from older vintages, and half came in the form of unrealized fair value gains as the recently declining UMI trend worked its way into our various fair value marks.
The volume of loan transactions across our platform was approximately 241,000, up 102% from the prior year but down 2% sequentially and representing 163,000 new borrowers. Average loan size of approximately $8,865 nudged up from $8,580 in the prior quarter as the proportion of loans made to Super Prime borrowers increased.
Our Contribution Margin, a non-GAAP metric which we define as revenue from fees, minus variable costs for borrower acquisition, verification and servicing, as a percentage of revenue from fees, came in at 55% in Q1, down 6 percentage points from the prior quarter and 2 points below guidance, largely due to the lower takes rates realized in the primer borrower segments where we exceeded expectations.
GAAP operating expenses were roughly $218 million in Q1, down 3% sequentially from Q4. Expenses that are considered variable, relating to borrower acquisition, verification and servicing, were up 7% sequentially relative to a 2% decline in transaction volume, which is reflective of the lower contribution margins previously referenced. Fixed expenses were down 8% quarter over quarter, as some of the temporary catchup accruals from Q4 rolled off in the new year.
Taken together, Q1 GAAP net loss was $2 million, well ahead of expectation and reflecting the outperformance on net interest income against our tightly managed fixed cost base. Adjusted EBITDA was $43 million, also scaling nicely in accordance with our operating leverage. Adjusted earnings per share was $0.30 based on a diluted weighted average share count of 104 million.
We ended Q1 with $726 million of loans held directly on the balance sheet, which is down 21% from the prior year but up sequentially from $703 million in Q4, as our newer R&D products continued to scale. Additionally, we recognized $89 million in loans from the consolidation of a securitization deal in which we retain minimal economic exposure.
We ended Q1 with unrestricted cash of approximately $600 million, down from $788 million in the prior quarter, with the delta primarily going towards R&D loan funding for new products, third-party risk capital partnerships, and a reduction in working capital from beginning-of-year corporate bonus payouts. In Q1, we put in place a universal shelf and a $500 million “At the Market” program, which gives us additional balance sheet flexibility.
We remain extremely pleased with our network of third-party capital relationships, and are excited to welcome the Fortress Investment Group as a new committed capital provider on our platform. Overall these relationships now comprise well north of 50% of the funding on our platform, and they are demonstrating their intended resilience in the current market climate of uncertainty.
As we look to Q2 and the remainder of 2025, much of the discussion has shifted to potential impacts from the macroenvironment. As mentioned, we’ve seen little discernible impact of the macro on credit performance so far. Uncertainty has increased, and we see the potential for both upside and downside scenarios that are credible in the near to medium term.
The main near-term risk in our estimation is reinflation, as any persistent tariffs placed on a significant share of our import economy will make things less affordable for consumers, a clear negative influence on credit. On the other hand, in the context of an already high-default environment with razor-thin savings rates, our continuing stance is that any dynamic which tempers current high levels of consumption back into line with underlying income would represent a positive effect on credit.
We remain sanguine about the prospective risks in the labor market, where we observe as many job openings in the economy as there are unemployed workers, and continue to perceive a structural shortage of labor supply that is under continuing pressure from demographic trends.
Given this context, our macro assumptions for the duration of this year remain consistent with the prior quarter, namely, we factor in no explicit expectation of any rate cuts and are planning for a steady macroenvironment in which the Upstart Macro Index remains in a range of 1.4-1.5. With this as background, for Q2, we are expecting:
- Total revenues of approximately $225 million, consisting of Revenue from Fees of approximately $210 million, and total net interest income of approximately positive $15 million
- Contribution Margin of approximately 55%
- Net income of approximately negative $10 million
- Adjusted net income of approximately positive $25 million
- Adjusted EBITDA of approximately $37 million with a
- Basic weighted average share count of approximately 96 million shares and a
- Diluted weighted average share count of approximately 104 million shares
For the full year of 2025, we now expect:
- Total Revenues of approximately $1.01 billion, consisting of Revenue from Fees of approximately $920 million, and net interest income of approximately positive $90 million
- Adjusted EBITDA margin of approximately 19%
- And we expect GAAP Net Income to be positive in the second half of the year and positive for the full calendar year
Beyond the numbers, I would just like to reiterate our gratitude to all of the hard-working teams at Upstart who make these results achievable. And with that, operator, over to you for Q&A.