UPST Upstart Holdings Stock Outlook 2026: Can AI Credit Scoring Survive Its Own Success?

Upstart Holdings is an uncomfortable stock to own. It lost 97% of its value from peak to trough in 2021–2022. Then it staged one of the most dramatic recoveries in fintech, growing originations 86% and revenue 64% in 2025 while returning to operating profitability. Now in 2026, management has doubled down with a national bank charter application — a move that could either cement Upstart’s moat or complicate its clean asset-light model.

This piece lays out what the verified numbers say, who the real competition is, and where the thesis breaks.

The Business: Fees, Not Balance Sheet

Upstart doesn’t want to be a bank — at least not yet. Its model is to be the “Intel Inside” for lenders: partner banks and credit unions originate loans using Upstart’s AI underwriting engine, and Upstart takes a fee. This keeps credit risk off Upstart’s books and lets it scale without raising expensive equity capital.

The revenue structure is worth understanding in detail. When a borrower is matched with a partner bank, Upstart earns an origination fee (typically a percentage of the loan amount), a referral fee if the borrower came through Upstart’s own channels, and a servicing fee for managing the loan over its life. A small fraction of loans that Upstart retains temporarily also generate net interest income, but this is a minor and declining portion of total revenue.

The AI model is Upstart’s core claim. It uses over 1,600 variables beyond traditional FICO scores — employment history, education, income trajectory, payment behavior across non-traditional data points — targeting “new prime” borrowers that legacy models reject. By the company’s account, this enables higher approval rates without higher default rates: a better deal for borrowers who get access to credit, and for banks that serve them.

This is either a genuine structural edge or a claim that hasn’t been stress-tested in a severe credit downturn. The 2022 period came close — Upstart’s own loan performance deteriorated when economic pressure hit subprime borrowers — but that cycle wasn’t as brutal as 2008. The question remains open.

What the data does confirm is that conversion rates have improved materially: from 15.1% in 2024 to 19.4% in 2025. More borrowers who apply are being approved and funded, which either means the model is getting better at identifying creditworthy applicants, or the risk bar has been lowered. Probably some of both.

Verified Financials: The Recovery Is Real

Full Year 2025 (Source: Upstart IR, February 2026):

• Total revenue: $1.0B, up 64% year-over-year
• Fee revenue: $950M, up 49%
• Loans originated: 1.497M, up 115%
• Total origination volume: ~$11B, up 86%
• Conversion rate: 19.4% (vs. 15.1% in 2024)
• Operating income: $42.6M (vs. –$173M in 2024)
• Auto and home loan originations each grew 5x

Q1 2026 (Source: Upstart 8-K, May 5, 2026):

• Total revenue: $308M (+44% YoY) — beat consensus
• Fee revenue: $277M (+49%)
• Loans originated: 425,356 (+77% YoY)
• Origination volume: ~$3.4B (+61% YoY)
• Home loans: +250% YoY, +16% QoQ; 25%+ fully automated; average close time 6 days
• Net loss: –$6.65M (widened vs. Q1 2025 –$2.44M)
• EPS: –$0.07, below consensus

2026 Guidance (reiterated Q1 call):

• Total revenues: ~$1.4B
• Fee revenues: ~$1.3B
• Adjusted EBITDA: ~$294M (~21% margin)

The EPS miss in Q1 was attributed to front-loaded hiring and infrastructure investment. Management flagged “more modest sequential growth” for Q2–Q4, implying they’re spending now to hit a $1.4B full-year target.

The Bank Charter: High Stakes Pivot

In March 2026, Upstart announced applications to the OCC and FDIC to form Upstart Bank, N.A., and simultaneously applied to the Federal Reserve to become a bank holding company.

The strategic logic is straightforward: retail deposits are cheap. If Upstart can fund a portion of its loans at 1–2% deposit rates rather than 6–8% institutional capital costs, the economics improve dramatically. It’s the same playbook SoFi ran in 2022, which stabilized its funding base during a period when capital markets froze.

The skeptic’s case: bank regulation is onerous. Capital ratios, liquidity requirements, and CRA obligations add cost and complexity. Existing partner banks may view Upstart as a competitor once it holds deposits. And approvals typically take two to three years — meaning this thesis plays out in 2028 at the earliest.

In the meantime, Upstart has diversified its institutional funding aggressively. Q1 2026 alone saw over $4B in new committed capital, including a Centerbridge deal ($1.2B), an Eltura/Aperture agreement ($1B), and a Wafra auto arrangement (~$200M). A $1B ABS issuance was multiple times oversubscribed. The 2022 scenario — where funding dried up and Upstart was forced to hold loans on its own balance sheet — is structurally less likely today.

Competitive Landscape

Three categories of competitors are worth watching, because they represent different attack vectors on Upstart’s model.

LendingClub (LC): A marketplace bank that holds loans on its own balance sheet, giving it direct skin in the credit game. LendingClub is less reliant on third-party capital than Upstart but bears direct credit losses when borrowers default. Its tech modernization has been steady but slower. Comparing UPST to fintech peers like AFRM illustrates how different business models produce very different risk profiles even in adjacent categories.

SoFi (SOFI): A digital bank targeting higher-FICO borrowers with mortgages, personal loans, credit cards, and investment accounts under one app. SoFi’s deposit base gives it structural funding advantages that Upstart currently lacks — it’s the single biggest reason why SoFi’s funding cost is more stable through rate cycles. The tradeoff: SoFi is a regulated bank, with capital requirements, CRA obligations, and regulatory oversight that constrain growth flexibility.

Pagaya (PGY): A B2B AI credit enhancement platform — smaller, lower-profile, but a direct competitor in selling AI underwriting infrastructure to lenders who don’t want to hand their workflow to Upstart.

Traditional banks themselves: This is the existential long-term question. JPMorgan Chase spent $17B on technology in 2024 alone. Wells Fargo and Bank of America are not standing still. If any major bank deploys an in-house AI credit model that matches Upstart’s accuracy — without sharing fees with a third party — Upstart’s partner bank network could erode. That’s probably a five to ten year threat, not a 2026 threat, but investors should think about it.

The most honest version of Upstart’s competitive moat is this: the AI model is a continuously trained system that improves with each loan originated on the platform. It has processed millions of loans across multiple credit cycles and product types. That dataset is hard to replicate quickly. But it’s not impossible.

Three Scenarios for 2026–2027

Bull case (30% probability): The Fed cuts rates twice more in 2026. Origination volume accelerates. The $1.4B revenue guide is achieved with upside. Bank charter approval timeline is fast-tracked. P/S re-rates to 8–10x on the new funding narrative, implying a stock price in the $60–80 range from current levels.

Base case (50% probability): Revenue hits $1.3–1.4B. Adjusted EBITDA margin holds around 20%. Bank charter takes until 2028. The stock drifts in the $35–55 range, reflecting steady growth without a catalyst to expand multiples.

Bear case (20% probability): A credit cycle turns. Partner defaults rise, institutional funders pull back. Revenue growth decelerates sharply. The bank charter is denied or delayed indefinitely. The stock revisits $15–25 — still far above 2022 lows but painful from current levels.

Valuation Check

At approximately $2.7B market cap against $1.17B TTM revenue, UPST trades at roughly 2.3x trailing sales. Against the $1.4B 2026 guidance, the forward P/S is under 2x. For a company growing fee revenue near 50% annually and targeting 21% EBITDA margins, that multiple looks compressed — unless the market is pricing in meaningful execution risk.

Historically, UPST has traded at P/S of 20–30x during the 2021 bubble. It has also traded at 1x during the post-rate-hike panic. Neither extreme reflects the actual business. A normalized growth multiple of 5–8x forward P/S — call it $7–11B in market cap — seems reasonable if 2026 guidance is delivered, implying meaningful upside from current prices.

Tax Considerations for US Investors

UPST pays no dividends, so there’s no dividend income to manage — your only tax event is the sale.

Long-term capital gains (held 12+ months): 0%, 15%, or 20% depending on your taxable income, plus the 3.8% net investment income tax if your MAGI exceeds $200K (single) or $250K (married filing jointly). At the 20% LTCG rate plus NIIT, the effective rate tops out at 23.8% — meaningfully lower than short-term rates for high earners.

Account type matters a lot with a volatile growth stock. UPST in a Roth IRA eliminates the capital gains question entirely on qualified withdrawals. In a traditional IRA or 401k, you defer taxes until withdrawal, but all gains are taxed as ordinary income — potentially worse than LTCG rates if the position grows significantly. For a tax-advantaged account with a long time horizon, that tradeoff may still favor holding here.

Tax-loss harvesting: UPST’s historical volatility creates frequent opportunities. If you’re sitting on a loss, harvesting it against realized gains elsewhere before year-end can be valuable. The 31-day wash-sale rule applies — you’d need to wait a month before re-entering the position, during which time the stock could move significantly in either direction.

Charitable giving: High-conviction investors who’ve held UPST from lower prices might consider donating appreciated shares directly to a donor-advised fund, avoiding capital gains tax entirely while getting the full fair-market-value deduction. This only works if you’ve held the shares over 12 months and the position is actually appreciated.

My Position: Watchlist, Not Wallet — Yet

I’m not initiating a full position at current prices. Here’s why: the EPS miss, the class action lawsuit with a June 8, 2026 lead plaintiff deadline, and the unresolved bank charter timeline all create near-term noise that doesn’t need to be in my portfolio right now. None of these are fatal individually, but together they add uncertainty over the next 60–90 days that I’d prefer to let resolve.

That said, I’m watching three specific triggers before I act:

1. The class action settles or proceeds without material financial exposure. Securities litigation involving AI model performance claims is the kind of thing that can either fizzle quickly or become a prolonged overhang. I want clarity.
2. Q2 2026 EPS at or above consensus. The Q1 miss was attributed to front-loaded investment. If Q2 shows EPS recovery, it validates that the spending was one-time rather than structural margin compression.
3. Any Fed rate cut signal — either a formal FOMC rate cut or a meaningful shift in the dot plot toward cuts. This improves origination economics and investor appetite for rate-sensitive growth stories simultaneously.

If two of three materialize, I’d enter at 3–5% of portfolio weight. I’d add on the bank charter news — that’s a slow-moving catalyst that doesn’t need to be in the price now but will be a significant re-rating event when (if) approved.

The forward P/S under 2x is genuinely cheap for 44%+ revenue growth if the growth rate holds. The problem is that UPST is one Fed hiking cycle away from a completely different financial reality. Position sizing and patience matter more than precision on this one.

The broader thesis — that AI credit scoring is structurally superior to FICO-based models and will take significant share from traditional bank underwriting over the next decade — is plausible and, if correct, makes the current price look cheap in hindsight. But thesis validation takes time, and the stock will be volatile throughout.

---

This article is for informational purposes only and does not constitute investment advice. All investment decisions are your own responsibility.

Verified sources: Upstart Holdings IR — Q4/FY2025 earnings release (February 2026); Q1 2026 8-K filing (May 5, 2026); Upstart national bank charter press release (March 2026); SEC EDGAR 10-Q FY2025. Market cap and price data reflect early June 2026 conditions and are subject to change.