Prediction: This Artificial Intelligence (AI) Stock Is Going to Double Before 2026 Is Over

Upstart reported 2025 revenue of $1.04 billion, up 64% year over year, with GAAP net income of $53.6 million versus a $128.5 million loss in 2024 and adjusted EBITDA of $230.4 million, up more than 20x. The company said its AI system handled 91% of Q4 2025 loan applications autonomously and it plans to apply for a national bank charter to launch an AI-powered bank. Shares are already down 40% this year and trade at 2.7x sales, or 1.8x based on 2026 revenue estimates, supporting the bullish case in the article.

Analysis

The market is still pricing Upstart like a cyclical originator with platform risk, not like a software-enabled underwriting utility with potential balance-sheet optionality. The important second-order effect is that a bank charter would change the company’s economics from pure fee capture to spread capture plus deposit funding, which—if executed without a credit hiccup—could materially expand revenue per application and reduce dependence on outside funding partners. That said, the charter path also imports a new regulatory regime, higher compliance costs, and a much lower tolerance for model drift, so the market is likely underestimating execution risk over the next 6-18 months.

The biggest competitive implication is not that Upstart “beats” FICO outright, but that it forces lenders to treat underwriting as a continuously retrainable data asset rather than a static score. That pressures incumbents and could compress pricing power across adjacent risk-scoring vendors and loan aggregators if lenders start demanding AI-driven lift metrics instead of legacy score thresholds. The more interesting second-order winner may be deposit-funding partners and loan buyers that can use Upstart’s funnel to source risk-adjusted yield, while the loser is any lender whose underwriting stack remains manual and slow.

The contrarian miss is that the bull case is increasingly a multiple story, not just a fundamentals story. A move from 2.7x to even the low end of its historical sales range would require sentiment to stay hot, but one quarter of slower originations, higher delinquencies, or delay in the charter approval process could re-rate the stock back toward distressed-fintech territory quickly. The right framework is months-to-years, not days: this is a long-duration optionality trade with acute event risk around credit performance and regulatory gating.

Near term, the setup looks better for upside calls than for outright stock chasing because implied volatility likely remains elevated while the chart is still healing from the drawdown. If the bank charter advances, the stock can gap on narrative expansion; if it stalls, the downside is likely driven by multiple compression rather than near-term earnings collapse. The asymmetric risk is that investors are paying for an AI-bank call option while ignoring that the underlying credit cycle could still deteriorate before the operating leverage fully shows up.

Key Decisions for Investors

• Initiate a starter long in UPST with a 6-12 month horizon; size it as an optionality position, not a core holding, because the upside is driven by multiple expansion + charter approval while downside is dominated by regulatory/credit event risk.
• Prefer a call spread in UPST over outright equity: buy 9-12 month call spreads to capture a potential re-rating if the bank-charter process advances, while limiting loss if approval is delayed or originations decelerate.
• Pair trade: long UPST / short FICO for 6-18 months as a thematic bet on AI-driven underwriting adoption; the trade works only if lenders increasingly reward dynamic models over legacy scoring, but cut quickly if Upstart credit metrics weaken.
• Avoid chasing FICO weakness solely on narrative; use any rally in FICO as a chance to own the incumbent beneficiary of slower adoption and higher lender conservatism.
• Set a risk trigger on UPST around any evidence of rising delinquencies, tighter funding availability, or charter delay; those are the catalysts that can compress the multiple faster than consensus expects.