Upstart: Riding the AI wave

Upstart is a leading financial technology company focused on transforming consumer lending through AI. By replacing traditional credit scoring models with a data-driven approach, the company connects borrowers with banks and institutional investors through a cloud-based marketplace that aims to expand access to affordable credit. With a growing network of lending partners, continuous improvements in its AI models, and expansion into new lending categories such as auto and home loans, Upstart is working to build a scalable and more efficient credit ecosystem. The question remains: Does this AI-driven disruptor deserve a spot in your portfolio?

This is not a financial advice. I am not a financial advisor and I only do these post in order to do my own analysis and elaborate about my decisions, especially for my copiers and followers. If you consider investing in any of the ideas I present, you should do your own research or contact a professional financial advisor, as all investing comes with a risk of losing money. You are also more than welcome to copy me. 

For full disclosure, I should mention that I do not own any shares in Upstart at the time of writing this analysis. If you would like to copy or view my portfolio, you can find instructions on how to do so here. If you want to purchase shares or fractional shares of Upstart, you can do so through eToro. eToro is a highly user-friendly platform that allows you to get started on investing with as little as $50.

The Business

Upstart Holdings, Inc. was founded in 2012 and operates a cloud-based artificial intelligence lending marketplace designed to improve how credit decisions are made. The company connects consumers with a network of more than 100 banks, credit unions, and institutional investors, enabling these partners to originate loans using Upstart’s proprietary AI models rather than relying primarily on traditional credit scoring systems such as FICO. While legacy models tend to focus heavily on a limited set of variables related to credit history, Upstart’s platform evaluates more than 2.500 data points, including education, employment history, income patterns, transaction behavior, and cost of living. This broader approach allows lending partners to more accurately assess risk, often resulting in higher approval rates and lower interest rates for borrowers without increasing default risk. Upstart operates an asset-light business model, acting as a technology provider rather than a traditional lender. Consumers can apply for loans through Upstart’s website, through partner-branded interfaces, or through integrations such as auto dealerships, and once approved, loans are funded by banks, credit unions, or institutional investors rather than by Upstart itself. The company earns the majority of its revenue through platform and referral fees when loans are originated through its system, as well as servicing fees for managing the loan lifecycle. In some cases, Upstart temporarily holds or packages loans before selling them to investors, including through securitization structures, which adds flexibility to its funding model and helps scale the platform. The business spans multiple credit categories, including personal loans, auto loans, and home equity lines of credit, and aims over time to become a broad marketplace for consumer lending. A key element of the platform is its high level of automation, with the vast majority of loans processed without human intervention, which reduces costs and creates a faster and more seamless experience for both borrowers and lenders. Upstart’s competitive moat is primarily built on its proprietary AI models, data advantage, and the flywheel effects created by its marketplace. Over more than a decade, the company has trained its models on tens of millions of repayment events, allowing it to continuously improve its ability to separate risk and price loans more accurately. Each additional loan originated on the platform generates new data that feeds back into the models, making them more precise over time and reinforcing a self-improving cycle. This creates a significant barrier to entry, as competing platforms would need both large-scale loan origination and years of performance data to reach similar levels of predictive accuracy. Beyond underwriting, Upstart applies its AI across the entire lending process, including fraud detection, income and identity verification, prepayment prediction, and servicing optimization, further strengthening its technological edge. The company also benefits from network effects within its ecosystem. As more borrowers use the platform, more lending partners and investors are attracted by the improved risk assessment and loan performance, which in turn increases funding availability and enhances the platform’s value for borrowers. This interconnected marketplace strengthens Upstart’s position over time. Switching costs also play a role in the moat, as financial institutions that integrate Upstart’s models into their lending processes and experience improved approval rates and efficiency face meaningful risk in reverting to less advanced systems. Additionally, the high level of automation reduces reliance on manual underwriting, creating cost advantages and supporting scalability as loan volumes grow. Finally, the complexity of building and maintaining high-performing AI lending models, which require both advanced machine learning expertise and large volumes of proprietary data, makes Upstart’s competitive position difficult to replicate. Together, these factors form a data-driven and technology-based moat that, if maintained, could allow Upstart to play an increasingly important role in modernizing the consumer credit ecosystem.

Management

Paul Gu is the CEO of Upstart, a role he assumed on May 1, 2026, after previously serving as the company’s co-founder and Chief Technology Officer. His appointment represents a high-continuity leadership transition, as he has helped build Upstart alongside co-founder Dave Girouard since the company was founded in 2012. Rather than bringing in an external executive, Upstart is moving from one founder-led chapter to another, with Dave Girouard continuing as Executive Chairman and special advisor. This gives the company continuity in strategy, culture, and long-term vision while placing a deeply technical founder at the center of its next phase. Paul Gu’s background is rooted in engineering and data science, which is highly relevant for a company whose entire business model depends on artificial intelligence. He studied computer science and economics at Yale University, where he graduated at a young age and was recognized for his strong analytical abilities. After completing his undergraduate studies, he pursued a PhD in computer science at Stanford University, focusing on machine learning and artificial intelligence, although he left the program to co-found Upstart. This combination of academic training in both economics and advanced machine learning has shaped his approach to building models that can assess credit risk more accurately than traditional systems. Before founding Upstart, Paul Gu also gained experience in the technology sector, including working at Google, where he was exposed to large-scale data systems and machine learning applications. This early experience helped inform the technical foundation of Upstart’s platform, which relies on analyzing thousands of variables to improve lending decisions. As CTO, he played a central role in developing the company’s AI models, data infrastructure, and product architecture, including expanding the platform from personal loans into areas such as auto lending and home equity lines of credit. The leadership transition to CEO also appears to have been carefully planned over several years. Dave Girouard stated that the companies had discussed this milestone for more than five years and described the transition as one of the most thoroughly prepared succession plans he could imagine. Paul Gu has emphasized that he spent more than 14 years working alongside Dave Girouard, helping build the company from its earliest days, including finding its first customers, securing initial funding partners, scaling the platform, taking the company public, and navigating the volatility of the COVID stimulus period. This matters because Upstart is still a relatively young and cyclical company, and its future depends on disciplined execution across technology, funding, and credit performance. The transition therefore does not signal a strategic shift, but rather a continuation of the existing direction under a founder who has been deeply involved in the most important parts of the business since inception. According to management, Paul Gu and Dave Girouard have effectively operated as aligned leaders and co-decision makers for years, and with Dave Girouard remaining closely involved as Executive Chairman and the company’s largest shareholder, the risk of disruption is limited. Given Paul Gu’s technical expertise, academic background in machine learning, and long-term involvement in building Upstart’s platform, he appears well positioned to lead the company through its next phase of growth. His main challenge will be proving that Upstart’s AI models can consistently deliver strong credit outcomes across cycles while scaling into larger lending markets and maintaining stable funding, which will ultimately determine whether the company can fulfill its ambition of transforming the consumer credit system.

The Numbers

The first number we will look into is the return on invested capital, also known as ROIC. We want to see a 10-year history, with all numbers exceeding 10% in each year. Upstart made its IPO in December 2020, so we only have a limited history as a public company, and the results have been far from what we would like to see. ROIC was strong in 2021 at around 12,5%, but then turned negative in 2022, 2023, and 2024 before recovering slightly to a positive level in 2025. This volatility is not random but closely tied to the nature of Upstart’s business model and the broader credit cycle. The most important driver behind the weak and volatile ROIC is Upstart’s sensitivity to macroeconomic conditions. When interest rates rise and credit conditions tighten, banks and institutional investors become more cautious and reduce their appetite for lending. This directly impacts Upstart because its revenue depends on loan origination volumes. When volumes decline, revenue falls quickly, while a large part of the cost base remains. This creates a situation where profitability drops sharply, which in turn drives ROIC lower and in some years into negative territory. Another key factor is the company’s decision to continue investing through the downturn. Upstart has kept spending on engineering, AI model development, and platform expansion even during periods of lower revenue. While this is rational from a long term perspective, as the company is trying to strengthen its competitive moat, it puts pressure on short term returns. In simple terms, the company is increasing its cost base and invested capital at a time when earnings are under pressure, which reduces ROIC. A third important driver is balance sheet usage. Although Upstart is designed to be an asset light platform, it has at times taken loans onto its own balance sheet when funding from partners and investors has been constrained. This increases invested capital significantly while also introducing more risk. Holding loans ties up capital and exposes the company to credit losses, which further depresses returns. Debt has also played a role. To support loan warehousing and operations during periods of weaker external funding, Upstart has taken on additional financing. This increases the capital base and interest costs, both of which weigh on ROIC. When combined with lower earnings, this creates a double impact that pushes returns down. It is also worth noting that Upstart’s business model has a degree of operating leverage. In good times, when loan volumes are high, the platform can generate strong margins because a large portion of costs are relatively fixed. This is what we saw in 2021. However, the same dynamic works in reverse during downturns, where declining volumes lead to a sharp drop in profitability. This explains why ROIC can swing so dramatically from positive to negative in a short period of time. Looking ahead, there are reasons to believe that ROIC can improve, but it is unlikely to become as stable as what we see in more mature or less cyclical businesses. If interest rates stabilize or decline and credit conditions improve, loan origination volumes should recover, which would lift revenue and profitability. At the same time, if Upstart can reduce its reliance on holding loans on its own balance sheet and return closer to a pure platform model, invested capital should decline relative to earnings, which would support higher ROIC. The company has already shown early signs of improvement in 2025 with a return to positive ROIC, suggesting that the business is starting to recover as conditions normalize. However, it is important to recognize that Upstart’s ROIC will likely remain more volatile than what we would ideally want. The company operates in a cyclical industry, depends on external funding markets, and continues to invest heavily in growth and technology. These factors make it difficult to achieve consistently high and stable returns on capital. The key question is therefore not whether ROIC will be perfectly stable, but whether it can trend upward over time as the platform scales, the AI models improve, and the business becomes less dependent on balance sheet usage. If Upstart can demonstrate this over the next several years, ROIC could become a more attractive part of the investment case.

The next numbers are the book value + dividend. In my old format this was known as the equity growth rate. It was the most important of the four growth rates I used to use in my analyses, which is why I will continue to use it moving forward. As you are used to see the numbers in percentage, I have decided to share both the numbers and the percentage growth year over year. To put it simply, equity is the part of the company that belongs to its shareholders – like the portion of a house you truly own after paying off part of the mortgage. Growing equity over time means the company is becoming more valuable for its owners. So, when we track book value plus dividends, we’re essentially looking at how much value is being built for shareholders year after year. Upstart’s equity development has been quite volatile since the IPO. Equity peaked in 2021 and then declined in 2022, 2023, and slightly again in 2024 before increasing meaningfully in 2025. This pattern reflects the same underlying dynamics that affected ROIC, namely the credit cycle, profitability, and the company’s use of its balance sheet. The most important driver behind the decline in equity in the years after 2021 has been net losses. When a company generates losses, retained earnings decrease, which directly reduces equity. Upstart experienced this during the period of rising interest rates and tighter credit conditions, where loan volumes declined and profitability came under pressure. Another key factor is the company’s increased balance sheet exposure. Although Upstart is designed to operate as a platform, it has at times held more loans on its own balance sheet when funding from partners and institutional investors was constrained. This ties up capital and increases risk, and if those loans are marked down or generate lower returns than expected, it can weigh on equity. Even without large losses, simply holding more assets relative to earnings can reduce the efficiency with which equity compounds over time. At the same time, Upstart has continued to invest in its platform, including its AI models, product expansion, and infrastructure. These investments are important for long term growth, but during periods of lower revenue they reduce profitability, which again feeds through to weaker equity development. Unlike more mature companies, Upstart is not returning capital to shareholders through dividends or buybacks, so changes in equity are primarily driven by earnings and balance sheet movements rather than capital allocation decisions. The improvement in 2025 is therefore an important signal. The increase in equity suggests that profitability has started to recover, or at least that losses have narrowed significantly, allowing retained earnings to stabilize and grow again. This aligns with the broader recovery in loan volumes and revenue that the company experienced as credit conditions began to normalize. In simple terms, the business is once again generating value instead of consuming it, which is a necessary condition for sustained equity growth. Looking ahead, equity can continue to grow, but it is unlikely to do so in a smooth and predictable manner. Upstart operates in a cyclical industry, and its performance is closely tied to interest rates, credit demand, and investor appetite for loans. In favorable environments, the company can scale quickly, generate profits, and grow equity at an attractive rate. In more challenging environments, the opposite can happen, leading to periods of stagnation or decline. Another important factor will be whether Upstart can move back toward a more asset light model. If the company relies less on holding loans on its own balance sheet and instead facilitates loans purely through partners and investors, it should be able to grow equity more efficiently over time.

Finally, we will analyze the free cash flow. Free cash flow, in short, refers to the cash that a company generates after covering its operating expenses and capital expenditures. I use levered free cash flow margin because I believe that margins offer a better understanding of the numbers. Free cash flow yield refers to the amount of free cash flow per share that a company is expected to generate in relation to its market value per share. Upstart’s free cash flow profile has been highly volatile since the IPO, with positive free cash flow in only two out of the five years available. The company generated strong free cash flow in 2021 and again in 2024, but this was followed by significant negative free cash flow in 2022, 2023, and again in 2025. This volatility is closely tied to the nature of Upstart’s business model and its exposure to the credit cycle. The most important driver of free cash flow volatility is loan origination volume and funding availability. Upstart earns the majority of its revenue from fees when loans are originated through its platform. When credit conditions are strong and funding from banks and institutional investors is readily available, loan volumes increase, revenue grows, and the company can generate significant cash. This is what we saw in 2021 and again during the recovery in 2024. However, when interest rates rise and investors become more cautious, funding tightens and loan volumes decline. This leads to lower revenue while costs do not fall at the same pace, which puts pressure on cash generation and can quickly turn free cash flow negative. Another major factor is balance sheet usage. Although Upstart is designed to operate as an asset light platform, it has at times held loans on its own balance sheet when external funding was constrained. This has a direct and often large impact on free cash flow because capital is tied up in loans rather than being returned as cash. When the company increases its loan holdings, it effectively uses cash to fund those loans, which reduces or even reverses free cash flow. This was a key reason for the large negative free cash flow in 2022 and 2023, and it also helps explain why free cash flow turned negative again in 2025 despite improving profitability. In simple terms, the business may be earning more on paper, but cash is being used to support loan growth. Operating leverage also plays an important role. Upstart has a relatively fixed cost base, particularly in engineering, data science, and platform infrastructure. When revenue grows quickly, as management highlighted with strong revenue growth and limited expense growth, a large portion of that incremental revenue can translate into profit and cash flow. This explains the sharp improvement in 2024. However, the same dynamic works in reverse when volumes decline or when the company reinvests heavily in growth initiatives, which can quickly reduce cash flow. It is also important to consider that Upstart continues to invest in expanding its platform, including new loan categories such as auto and home equity, as well as improving its AI models and infrastructure. These investments are necessary to strengthen the company’s long term competitive position, but they also consume cash in the short term. As a result, free cash flow should not be viewed in isolation without considering where the company is in its investment cycle. The return to negative free cash flow in 2025 is therefore not necessarily a sign of a deteriorating business, but rather a reflection of continued growth, balance sheet usage, and reinvestment. Management has indicated that revenue growth is outpacing expense growth and that profitability is improving, which suggests that the underlying earnings power of the business is strengthening. If the company can maintain this trend while reducing its reliance on holding loans on its own balance sheet, free cash flow should improve over time. Looking ahead, free cash flow is expected to become positive again, but it will likely remain volatile. The key drivers will be the health of the credit market, the availability of external funding, and how much of the loan volume Upstart chooses or needs to hold on its own balance sheet. In a favorable environment where funding is abundant and the company operates closer to a pure marketplace model, free cash flow could be strong due to the platform’s operating leverage and relatively low capital requirements. In a more constrained environment, free cash flow may again turn negative as the company steps in to support loan origination. Upstart primarily uses its free cash flow to reinvest in the business rather than returning capital to shareholders. This includes investments in AI model development, product expansion, and platform infrastructure, all of which are critical to maintaining and strengthening its competitive moat. In addition, the company uses capital to support its funding ecosystem, including temporarily holding loans, structuring securitizations, and building relationships with institutional investors. These uses of cash are aimed at enabling long term growth and making the platform more resilient across credit cycles. Over time, if the business matures and becomes more consistently cash generative, the company may have the option to allocate capital differently, but for now the priority remains reinvestment and scaling the platform.

Debt

Another important area to investigate is debt, and we want to see whether a business has a reasonable level of debt that could be paid off within three years. To assess this, we divide total long-term debt by earnings. When applying this measure to Upstart, the result shows that it would take approximately 42,6 years of earnings to pay off its long-term debt, which is far above the three-year threshold. However, this number should be interpreted with caution. The calculation is based on very low earnings in 2025, which significantly distorts the result. If we instead use earnings from 2021, which was the last year with strong profitability, the ratio improves to around 19,2 years. This is still high, but it highlights how sensitive this metric is to fluctuations in earnings. It is also important to understand the nature of Upstart’s debt. Unlike many traditional companies, a portion of Upstart’s debt is tied to its lending activities rather than its core operating business. When the company holds loans on its balance sheet, it often finances those loans using debt. This means that part of the debt is effectively matched by loan assets, making it different from debt used purely to fund operations or acquisitions. As a result, the headline debt number can appear more concerning than it actually is if viewed in isolation. That said, the increase in debt since the IPO is still worth paying attention to. The rise has largely been driven by periods where funding from banks and institutional investors was constrained, which led Upstart to step in and hold more loans itself in order to keep the platform running. This increases both financial risk and capital intensity, and it is one of the reasons why returns on capital and free cash flow have been under pressure in recent years. Management has indicated that the long-term goal is to operate closer to an asset-light marketplace model, where loans are primarily funded by partners rather than held on Upstart’s own balance sheet. If the company succeeds in this, both debt levels and balance sheet risk should decline over time. This would make the business more predictable and improve overall financial quality. It is also encouraging that Upstart has taken steps to manage its debt profile. By refinancing its convertible debt and extending maturities to 2029 and 2030, the company has reduced near-term refinancing risk and given itself more time to grow into its capital structure. This lowers the risk of financial stress in the short term. In summary, I believe Upstart’s debt position is a concern, but not necessarily a structural problem if viewed in the right context. The key issue is not just the absolute level of debt, but how much of it is tied to holding loans on the balance sheet. If Upstart can reduce this exposure and return to a more asset-light model while improving profitability, the debt metrics should improve significantly. Until then, the company’s debt levels should be monitored closely, especially in relation to earnings, loan holdings, and overall credit market conditions.

Unlock Exclusive Seeking Alpha Discounts – Level Up Your Investing With Zero Risk

If you’ve been thinking about improving your investing process, this is the easiest way to start. These offers are only available through my links, and the Premium plan even comes with a 100% risk-free 7-day trial. Try everything for a week, and if it’s not for you, just cancel. You lose nothing.

1) Seeking Alpha Premium — Try It Free for 7 Days

Access the tools I personally use every day:

• Earnings transcripts

• Stock screeners

• Deep-dive analysis

• Portfolio tracking

• Market news with context that actually matters

Special Price: $269/year (normally $299) + 7-day free trial (for new users only)

Try Premium Free for 7 Days → HERE

(Explore everything — cancel anytime during the trial and pay $0.)

2) Alpha Picks — Proven Stock Ideas

This stock-picking service has delivered +287% returns vs. the S&P 500’s +77% (July 2022–Nov 2025).Great for investors who want curated, long-term picks backed by data.

Special Price: $449/year (normally $499)

Get Alpha Picks → HERE

(Although Alpha Picks doesn’t offer a free trial, its historical outperformance means the subscription can often pay for itself quickly if results persist. For many investors, the potential return far outweighs the upfront cost).

3) Premium + Alpha Picks Bundle — Best Value

Get both services together and save $159.Perfect if you want both broad tools and high-conviction stock ideas.

Special Price: $639/year (normally $798)

Get the Bundle → HERE

(This bundle doesn’t include a free trial, but it gives you both services at a $159 discount. You get Premium’s in-depth research plus Alpha Picks’ high-performing recommendations, making it the most comprehensive option for serious investors.)

Risks

Macroeconomic factors is a risk for Upstart because the company operates at the intersection of consumer credit demand and the supply of funding from banks and institutional investors, both of which are highly sensitive to the broader economy. When economic conditions weaken due to rising interest rates, inflation, or declining consumer confidence, both sides of Upstart’s marketplace can be affected at the same time. Borrowers may delay taking out loans or become less qualified due to lower income, higher living costs, or job uncertainty, while lending partners and investors may reduce their appetite for funding loans due to increased risk or tighter liquidity conditions. Because Upstart earns the majority of its revenue from fees on loan originations, any slowdown in loan volume can have a direct and immediate impact on revenue and profitability. Interest rates play a particularly important role. When central banks keep rates high for an extended period, borrowing becomes more expensive for consumers, which reduces demand for loans. At the same time, higher rates increase the required returns for investors, making them more selective and cautious when allocating capital. This combination can significantly reduce the flow of loans through Upstart’s platform. In a more severe scenario, often referred to as a credit crunch, funding can dry up altogether, forcing Upstart to either reduce originations or step in and hold more loans on its own balance sheet. This increases risk and ties up capital, which can negatively impact both free cash flow and returns on capital. Macroeconomic conditions also affect credit performance. During economic downturns, unemployment may rise and household finances may come under pressure, which increases the likelihood that borrowers will miss payments or default on their loans. This is particularly relevant for Upstart because many of its borrowers have limited or non-traditional credit histories. These borrowers are often more vulnerable in difficult economic environments and may be more likely to prioritize essential expenses or secured obligations, such as mortgages or car payments, over unsecured personal loans. Higher default rates can reduce investor confidence in Upstart-powered loans, leading to lower funding availability and further pressure on the platform. Another important factor is the potential mismatch between model expectations and real-world outcomes during rapid economic changes. Upstart’s AI models are designed to assess credit risk based on historical data, but if macroeconomic conditions change quickly or behave differently than in the past, actual loan performance may deviate from model predictions. If defaults rise more than expected, investors may pull back from the platform, creating a negative feedback loop where reduced funding leads to lower originations, which in turn reduces data flow and revenue. This dynamic was visible during previous periods of market stress and highlights the sensitivity of the business to external conditions.

Funding availability is a risk for Upstart because the company depends on a continuous and diverse supply of capital from banks, credit unions, and institutional investors to fund loans on its marketplace. Unlike traditional lenders, Upstart does not primarily use its own balance sheet to fund loans, which means its ability to grow is directly tied to how much external capital is willing to flow through its platform. If this funding becomes constrained, loan originations decline, and because Upstart earns the majority of its revenue from fees on those originations, revenue and profitability can be impacted quickly. This makes the business highly dependent not only on borrower demand, but also on the willingness of third parties to provide capital across different market conditions. A key challenge is that this funding is not fully within Upstart’s control. Institutional investors and lending partners continuously adjust their activity based on market conditions, risk appetite, and expected returns. In periods of uncertainty, weaker credit performance, or tighter financial conditions, investors may reduce or pause their purchases of loans. This can create a situation where borrower demand remains relatively stable, but there is not enough capital available to fund those loans. In such cases, Upstart may be forced to slow down originations, which directly reduces revenue, or step in and hold more loans on its own balance sheet. Holding loans increases financial risk and ties up capital, which can pressure both free cash flow and returns on capital. Another important factor is concentration risk. A relatively small number of lending partners account for a significant share of loan originations and revenue. If one or more of these partners decide to reduce their activity, whether due to internal priorities, regulatory pressure, or changes in risk tolerance, it can have a meaningful impact on the platform. Because onboarding new partners requires time, integration, and regulatory alignment, it is not always possible to replace lost funding quickly. This creates a potential gap where loan demand cannot be matched with sufficient capital, limiting growth and increasing volatility in results. Counterparty risk also plays a role. Upstart relies on a network of financial institutions and investors to participate in its ecosystem, and the stability of this network is essential for maintaining consistent funding. If one or more counterparties fail to meet their obligations due to financial distress or market disruption, it can lead to losses or interruptions in funding. Even temporary disruptions can affect loan origination volumes and reduce confidence in the platform.

Laws and regulations is a risk for Upstart because the company operates in one of the most heavily regulated industries while relying on relatively new technologies such as artificial intelligence and a bank partnership model. This combination creates a complex and evolving regulatory environment where rules are not always clearly defined and are often subject to interpretation. Upstart must comply with a wide range of federal, state, and local laws covering areas such as consumer protection, fair lending, credit reporting, data privacy, and loan servicing. Any failure, or even perceived failure, to comply with these rules can lead to fines, penalties, restrictions on operations, or reputational damage. Because the company’s business model is built around facilitating loans rather than issuing them directly, it also depends on its lending partners remaining compliant, which adds another layer of regulatory exposure. A key challenge is that many of the existing regulations were not designed with AI-driven lending in mind. Upstart’s models use thousands of variables to assess creditworthiness, which can make it difficult to fully explain individual lending decisions in a simple way. Regulators are increasingly focused on issues such as transparency, explainability, and fairness in algorithmic decision making. If regulators determine that Upstart’s models result in unintended bias or “disparate impact” on protected groups, the company could face enforcement actions, be required to change its models, or even limit the use of certain data. This could reduce the effectiveness of its AI models and weaken its competitive advantage. Regulatory scrutiny is also increasing at the federal level. Agencies such as the Consumer Financial Protection Bureau have broad authority to oversee consumer lending practices and have shown a growing interest in fintech companies and AI-based decision making. These regulators can impose rules around what constitutes fair lending, acceptable fees, and borrower protections. They can also require companies to provide detailed documentation, testing, and monitoring of their models. Compliance with these requirements is both costly and time-consuming, and changes in regulatory priorities can lead to sudden shifts in what is considered acceptable practice. At the state level, regulation adds further complexity. Some states are questioning the structure of bank-fintech partnerships, particularly around who is considered the “true lender.” If regulators determine that the fintech rather than the bank is effectively issuing the loan, Upstart could be subject to stricter licensing requirements, interest rate caps, or other limitations. This could force the company to modify its marketplace model, limit its ability to operate in certain states, or reduce the attractiveness of its platform for lending partners. Because Upstart operates across many states, it must navigate a patchwork of different rules, which increases operational complexity and compliance costs. Legal and regulatory actions themselves also pose a risk. Even if claims are ultimately resolved in Upstart’s favor, investigations, lawsuits, or enforcement actions can be expensive, time-consuming, and disruptive. They can divert management attention, increase legal costs, and create uncertainty for investors, lending partners, and borrowers. In some cases, penalties can be significant, especially if they are calculated on a per-loan or per-violation basis. In addition, regulatory findings could require Upstart to compensate borrowers, change its practices, or restrict certain activities, all of which could negatively impact financial performance.

Reasons to invest

The ongoing innovation in its AI models is a reason to invest in Upstart because the company’s entire business model depends on its ability to assess credit risk more accurately than traditional methods. Unlike legacy credit scoring systems that rely on a limited set of variables, Upstart continuously improves its models by incorporating more data, refining its algorithms, and expanding its understanding of borrower behavior. As these models improve, they allow lending partners to approve more borrowers at lower interest rates while maintaining or even improving credit performance. This creates a powerful value proposition that can attract more borrowers, more lenders, and more capital to the platform over time. One of the most important aspects of this innovation is the scale and growth of Upstart’s proprietary data. The company has now trained its models on more than 100 million repayment events, built over more than a decade of operating history. This data is not publicly available and cannot simply be replicated by competitors, as it must be generated through real lending activity over time. As more loans are originated and repaid, the models continue to learn and improve, creating a self-reinforcing cycle where better models lead to better outcomes, which in turn attract more volume and generate even more data. Management has highlighted that the next 100 million repayment events are expected to come much faster, which could accelerate this improvement further. Innovation is also visible in the continuous rollout of new model versions and capabilities. Recent updates, such as Model 24 and Model 25, have improved the company’s ability to differentiate between higher and lower risk borrowers, a concept referred to as “risk separation.” These updates have expanded the training dataset by incorporating new types of borrowers and external data, allowing the models to better evaluate applicants who were previously difficult to assess. Improvements in accuracy, even by small percentages, can have a meaningful impact at scale, as they enable more precise pricing and reduce losses across millions of loans. Another important area of innovation is automation. Upstart is increasingly applying AI not only to underwriting, but also to other parts of the lending process such as verification, fraud detection, and servicing. For example, improvements in verification models have reduced default rates, while new AI tools, including voice-based systems, are helping automate processes that previously required manual intervention. Higher levels of automation reduce operating costs, improve the customer experience, and allow the platform to scale more efficiently. Over time, this can lead to stronger margins and higher returns on capital. Upstart also benefits from broader advancements in artificial intelligence. While its core lending models are proprietary, the company can leverage improvements in computing power, infrastructure, and base algorithms that are being developed across the wider AI ecosystem. This allows Upstart to build more complex and capable models without having to develop every underlying technology from scratch. At the same time, its specific application of AI to credit underwriting remains highly specialized, requiring unique data and tailored algorithms that are difficult for new entrants to replicate. Importantly, the company has a meaningful first-mover advantage. Upstart has been applying machine learning to lending for more than a decade, long before AI became a mainstream focus. Because the training data needed to build high-performing credit models must be generated through actual loan performance, early entry into the market provides a significant advantage. New competitors would need years of data collection and model refinement to reach a similar level of accuracy. This creates a technological moat that can strengthen over time as Upstart continues to innovate.

Newer lending products is a reason to invest in Upstart because they show that the company’s AI lending platform can expand beyond its original personal loan business and enter much larger credit markets. Upstart started primarily with unsecured personal loans, but it is now scaling products such as auto loans, home equity lines of credit, and small dollar loans. This matters because personal loans are only one part of the consumer credit market, while auto lending and home lending represent significantly larger opportunities. If Upstart can apply the same AI-driven underwriting, automation, and borrower experience to these categories, the company’s long-term addressable market could become much larger than it was at the time of its IPO. Management has indicated that auto and home originations both grew around five times year over year in 2025, which suggests that the newer products are beginning to find product-market fit rather than remaining small experiments. Auto lending is the most mature of these newer categories. Upstart has expanded both auto refinancing and dealership-based auto finance, and management noted that it has significantly increased the number of active dealership partnerships over time. This is important because dealership financing gives Upstart access to borrowers at the point of sale, where many car loans are originated. Once the company is integrated into a dealership, the cost of acquiring each additional loan can be relatively low, which could make the model more attractive as it scales. Upstart has also improved the auto buying process by enabling remote contract signing and using AI to automate document verification, reducing funding times and improving the customer experience. Home lending is another promising opportunity, particularly through HELOCs. These loans are generally larger than personal loans, and many existing Upstart customers may be homeowners who could benefit from using a HELOC to refinance debt or fund projects at lower rates. Management has highlighted that the company is using property data and cross-selling to its existing member base to grow this product. This could be an efficient acquisition channel because Upstart already has borrower relationships and data from its personal loan business. The company has also made progress in automating the HELOC process, reducing what has historically been a slow and paperwork-heavy product into something closer to a digital lending experience. Management has noted that Upstart’s HELOC process is significantly faster than the industry average, which could become a meaningful advantage if customers value speed, simplicity, and better pricing. Another encouraging sign is that the newer products are increasingly being funded by third parties rather than Upstart’s own balance sheet. This is important because one of the risks with launching new lending products is that the company may need to hold loans itself until investors and lending partners become comfortable with the credit performance. Management has stated that a large share of auto and home production is now funded by external partners, and it expects this trend to continue. If Upstart can scale these products while keeping most loans off its balance sheet, it supports the long-term thesis that the business can grow in a more asset-light and capital-efficient way. The economics of these products are also attractive, even though they may differ from personal loans. Auto loans and HELOCs may have lower upfront take rates, but they often have much larger average loan sizes and can generate servicing revenue over time. Management expects secured products in auto and home to contribute more than $100 million in fee revenue in 2026, which would make them a more meaningful part of the business. Over time, these products could also attract larger banks and institutional investors because secured credit categories may be more familiar and appealing to certain capital providers than unsecured personal loans.

Growing volume in personal loans is a reason to invest in Upstart because it demonstrates that the company is gaining market share in its core and most established product while improving profitability and efficiency. Personal loans remain the foundation of Upstart’s business, and strong growth in this category shows that its AI-driven platform is resonating with both borrowers and lending partners. Management has highlighted that originations in personal loans grew significantly year over year, which not only drives higher revenue but also confirms that Upstart is becoming more competitive in terms of pricing, approval rates, and user experience. Because the company earns fees on each loan originated, higher volumes translate directly into revenue growth and improved operating leverage when costs grow at a slower pace. One of the most important aspects of this growth is that it is driven by market share gains rather than relying solely on overall market expansion. Upstart’s product is increasingly competing on both best rates and best process, meaning borrowers can access credit more easily and often at better terms than through traditional lenders. As the company’s AI models improve, it can approve more borrowers without increasing risk, which allows it to capture a larger portion of the market. This creates a compounding effect where better performance leads to more volume, which in turn generates more data and further strengthens the models. Over time, this dynamic can help Upstart widen its lead over competitors and solidify its position in the personal loan market. Another key driver is the quality and consistency of loan performance. Upstart has demonstrated that its loan vintages can deliver attractive returns for capital providers, which is critical for maintaining and expanding funding on the platform. When investors and lending partners see consistent returns that exceed benchmarks such as government bonds, they are more likely to allocate additional capital to the platform. This increased funding availability supports further growth in loan originations. In this way, strong credit performance reinforces volume growth by attracting more capital, which is essential for scaling the business. The growth in personal loans also highlights the scalability of Upstart’s model. The company has shown that it can grow originations and revenue at a much faster rate than its cost base, particularly headcount. This indicates strong operating leverage, where incremental volume can be processed with relatively limited additional costs. As a result, higher loan volumes can lead to disproportionately higher profitability over time. This is an important characteristic for a platform business, as it suggests that margins can expand as the company scales. Another important factor is customer lifetime value. As Upstart grows its personal loan volume, it is also building a larger base of customers that it can engage with over time. Many borrowers return for additional loans or use other products offered by the platform. Management has indicated that this area is still relatively early, with significant potential to increase engagement and cross-sell opportunities. This means that the value of acquiring a new customer today may extend far beyond a single loan, supporting long-term growth and profitability. The expansion into higher-quality borrower segments also supports growth. Upstart is increasingly serving more prime and super-prime borrowers, who tend to have larger loan sizes and lower risk profiles. While these segments may have slightly lower fees, they can still generate attractive economics due to larger loan balances and better credit performance. At the same time, expanding into these segments broadens the addressable market and strengthens the platform’s appeal to a wider range of borrowers and capital providers.

Support the Blog

I want to keep the blog free and accessible for everyone. If you enjoy the content and would like to support it, you can buy me a cup of coffee through PayPal. Every little bit helps and is truly appreciated!

Valuation

Now it is time to calculate the share price. I perform three different calculations that I learned at a Phil Town seminar. If you want to make the calculations yourself for this or other stocks, you can do so through the tools page on my website, where you have access to all three calculators for free.

The first is called the Margin of Safety price, which is calculated based on earnings per share (EPS), estimated future EPS growth, and an estimated future price-to-earnings (P/E) ratio. The minimum acceptable rate of return is set at 15%. I chose to use an EPS of 0,45 from the year 2025. I have selected a projected future EPS growth rate of 15%. Finbox expects EPS to grow by 45,7% over the next five years, but 15% is the highest number I use. Additionally, I have selected a projected future P/E ratio of 30, which is double the growth rate. This decision is based on Upstart's historically higher price-to-earnings (P/E) ratio. Finally, our minimum acceptable rate of return has already been established at 15%. After performing the calculations, we determined the sticker price (also known as fair value or intrinsic value) to be $13,50. We want to have a margin of safety of 50%, so we will divide it by 2. This means that we want to buy Upstart at a price of $6,75 (or lower, obviously) if we use the Margin of Safety price.

The second calculation is known as the Ten Cap price. The rate of return that a company owner, or stockholder, receives on the purchase price of the company essentially represents its return on investment. However, I cannot calculate this for Upstart, as the company reported negative cash from operations in 2025.

The final calculation is referred to as the Payback Time price. It is a calculation based on free cash flow per share. However, I cannot calculate this for Upstart, as the company reported negative free cash flow in 2025.

Conclusion

I believe that Upstart is an intriguing company with strong management. The company has built a moat through its proprietary AI models, data advantage, and the flywheel effects created by its marketplace. Both ROIC and free cash flow have been volatile over the years, with several years of negative performance. This volatility is likely to continue in the near term, although it is expected to improve over the long term. Macroeconomic factors are a risk for Upstart because both borrower demand and investor funding are highly sensitive to the economic environment. When conditions weaken, loan volumes can fall, defaults can rise, and funding can tighten, creating a negative feedback loop that directly pressures revenue, profitability, and returns on capital. Funding availability is also a risk because the company depends on a steady supply of external capital from banks and investors to fund loans on its platform. If this funding slows or becomes more selective, loan originations can decline quickly, which directly reduces revenue and may force Upstart to take on more risk by holding loans on its own balance sheet. Laws and regulations represent another risk, as the company operates in a heavily regulated industry with an AI-driven model that faces evolving and sometimes unclear rules. Increased scrutiny around fair lending, transparency, and bank partnerships could lead to higher compliance costs, operational restrictions, or forced changes to its business model, which may impact growth and profitability. The ongoing innovation in its AI models is a reason to invest in Upstart because it enables the company to assess credit risk more accurately, approve more borrowers at better rates, and improve loan performance over time. As its models continuously learn from growing proprietary data, this creates a self-reinforcing advantage that attracts more borrowers, lenders, and capital while strengthening its competitive moat. Newer lending products are also a reason to invest because they show the platform can expand beyond personal loans into much larger markets such as auto and home lending, significantly increasing its long-term growth potential. As these products gain traction and attract third-party funding, they can diversify the business and support a more scalable and capital-efficient model. Growing volume in personal loans is another positive factor because it shows the company is gaining market share in its core business while improving efficiency and profitability. As higher volumes drive more revenue, better data, and stronger funding relationships, they create a compounding effect that can support long-term growth and scalability. While I believe there are things to like about Upstart, I also believe there are better opportunities in the market, and therefore I will not be investing in Upstart at this time.

My personal goal with investing is financial freedom. It also means that to obtain that, I do different things to build my wealth. If you have some extra hours to spare each month, you can turn a few hours a week into a substantial amount of money in a few years. If you are interested to know how I do it, you can read this post.

I hope you enjoyed my analysis! While I can’t post about every company I analyze, you can stay updated on my trades by following me on Twitter. I share real-time updates whenever I buy or sell, so if you’re making your own investment decisions, be sure to follow along!

Some of the greatest investors in the world believe in karma, and to receive, you will have to give (Warren Buffett and Mohnish Pabrai are great examples). If you appreciated my analysis and want to get some good karma, I would kindly ask you to donate a bit to dog meat rescue. Every year 30 million dogs are slaughtered and eaten by humans around the world! If you have enjoyed the analysis and want some good karma, I hope that you will donate a little to save these dogs here. Even a little will make a huge difference to save these wonderful animals. Thank you.