Prediction: 2026 Will Be the Year of Upstart

Upstart’s AI-driven lending platform has driven strong momentum—revenue was up nearly 80% through the first three quarters of 2025 and the company has returned to profitability—while its model reportedly enables 43% more loan approvals with no additional defaults and serves over 100 lender partners. A November signal from its algorithm trimmed Q3 originations by anticipating an economic headwind, which temporarily pressured growth, but analysts still expect roughly $2.38 in EPS for the year and the stock trades below 20x that estimate. Management guidance and analyst forecasts point to continued progress through 2027, and the firm argues that a full year of strong results should reveal the long‑term marketability of its credit‑scoring technology, potentially catalyzing a re-rating in 2026.

Analysis

Winners will be fintechs and bank partners that adopt higher‑quality ML scoring (Upstart/UPST and reseller partners), while legacy bureaus (EFX, TRU) face margin pressure on commoditized scores and slower product monetization. Upstart’s algorithm reducing originations today increases its pricing power long term: lenders will pay more for precision in a weak credit cycle, shrinking ABS issuance (lower supply) and concentrating demand for high‑quality underwriting tools. Cross‑asset: a demonstrable fall in originations can tighten ABS supply and lift spreads short term, while lower expected loss curves should marginally tighten bank funding spreads and reduce CYD for consumer credit bonds over 12–24 months.

Tail risks include regulatory intervention on opaque AI scoring, model breakdown in an unforeseen macro shock, or key bank partners building in‑house models; any of these could halve revenue growth in 12–18 months. Immediate (days) — sentiment/IV swings; short term (weeks–months) — guidance and quarterly validations will drive re‑ratings; long term (quarters–years) — recurring SaaS licensing and proof across a full credit cycle drives durable TAM capture. Hidden dependencies: revenue still scales with origination volume and bank balance‑sheet appetite, so sustained recession may compress bookings even if model quality is proven. Catalysts: four consecutive quarters of outperformance, a marquee national bank win, or an ABS shelf deal tied to Upstart originations.

Trade implications: tactical long UPST exposure via phased equity buys and concentrated LEAP calls captures asymmetric upside if model validation continues; fund a long UPST/short EFX pair to express secular share shift with lower market beta. Options: prefer 12–24 month LEAPs (buy ATM or 10–20% OTM) or buy call spreads to limit premium; avoid selling uncovered premium into uncertain event cadence. Sector rotate 2–4% from legacy bureaus into fintech SaaS and ABS long exposure; enter in tranches over 6–12 weeks keyed to quarterly results.

Contrarian view: the market misreads the November “warning” as pure demand risk when it is product validation — the algorithm doing precisely what it should — implying the downside reaction is likely overdone. Historical parallels: early FICO and automated underwriting adoption showed multi‑year lags between proof and valuation re‑rating; Upstart may follow a similar path but faster given AI tailwinds. Unintended consequences: broader regulatory scrutiny could force explainability requirements that erode Upstart’s edge, creating binary outcomes and asymmetric option‑like payoffs for holders.

Key Decisions for Investors

• Establish a phased long position in UPST equal to 2–3% of portfolio over 6–8 weeks (three equal tranches). Add another 2% if Upstart reports 4 consecutive quarters of revenue growth >30% and NTM EPS revisions are +20% within 12 months. Set initial stop‑loss at -30% from average entry or trim to half size if quarterly originations miss consensus by >15%.
• Implement a relative value pair: long UPST (notional 2%) versus short EFX (notional 1–1.5%) to express AI score adoption vs legacy bureaus; rebalance or close if the spread narrows by 20% in 3 months or widens by 40% in 6 months.
• Buy 12–24 month LEAP calls on UPST sized to 1–2% portfolio notional (prefer ATM or 10–20% OTM) or buy a call spread (buy LEAP ATM, sell 25–35% OTM) to cap cost. Target exit: 100% gain on option or if implied volatility collapses >40% without fundamental deterioration.
• Reduce direct exposure to legacy credit‑data providers (EFX/TRU) by 1–2% and redeploy into fintech SaaS or bank risk‑management software names over the next 3 months, prioritizing companies with recurring revenue and contractually sticky bank relationships.
• Monitor regulatory signals closely: if within 90 days a major regulator issues draft rules restricting opaque ML scoring or requires forced human‑explainability, reduce UPST exposure by 50% within 10 trading days and hedge with short fintech ETFs or buy protection (puts) on UPST.