I Recently Bought an Artificial Intelligence (AI) Stock That I Predict Will Double by the End of 2026

Upstart reported record Q1 revenue of $308 million, up 44% year over year, and $40 million in adjusted EBITDA, while originating 425,356 loans, up 77%. Management guided to $1.4 billion in 2026 revenue and sees 35% compound annual growth through 2028, implying more than $2.5 billion in revenue by then. The stock is down 33% year to date but trades at 2.6x trailing sales versus a 5.7x three-year average, supporting the author's view that shares could double.

Analysis

UPST is increasingly a model of software-like operating leverage inside a balance-sheet-light financial utility, but the key second-order effect is not just faster loan decisions — it is better unit economics for funding partners. If its approval engine keeps improving, the true moat shifts from underwriting accuracy to distribution efficiency and partner retention, which should compress acquisition costs and raise take rates over time. That makes the upside path less about loan growth alone and more about the market assigning a premium multiple to a repeatable underwriting infrastructure layer.

The biggest near-term risk is not competition from another fintech, but macro-driven cohort deterioration that can lag by several quarters. In a softer consumer environment, a model that looks superior on approval speed can still face higher charge-offs once vintages season, and funding partners can quickly demand tighter terms or reduced capacity. That creates a non-linear downside: revenue can keep growing while implied risk-adjusted economics deteriorate, which is exactly when sentiment can reverse sharply.

FICO is the obvious strategic loser in narrative terms, but the more relevant beneficiary set may include bank and captive funding partners that get incremental origination flow without building their own tech stack. The market is still treating this as a cyclical fintech rather than a platform transition, which may be why the multiple gap persists. However, that discount is only justified if investors believe model performance is brittle; if credit performance remains stable through the next consumer downturn, re-rating could happen faster than fundamentals because the story is already known and under-owned.

Consensus is probably underestimating how much of the upside is already embedded in growth expectations, while overestimating how quickly AI can expand across every loan type with equal economics. The asymmetry is that the stock can rerate on continued execution, but the downside is concentrated in one bad vintage or partner pullback. This is a classic ‘great technology, fragile timing’ setup, which argues for owning convexity rather than chasing common equity outright.

Key Decisions for Investors

• Long UPST via 3-6 month call spreads financed against out-of-the-money puts; target a 2:1 payoff if the market re-rates on continued originations, but cap downside if consumer data weakens.
• Pair trade: long UPST / short FICO for 2-4 quarters as a thesis on AI underwriting displacing legacy scoring; stop if FICO prints evidence of renewed enterprise share gains or UPST funding partners tighten materially.
• Add a tactical short-dated put hedge on UPST ahead of the next vintage/credit update; the setup offers asymmetric downside if delinquency trends lag higher even as headline revenue stays strong.
• Monitor bank-partner disclosures and securitization spreads over the next 1-2 earnings cycles; if funding costs widen, reduce UPST exposure because the market will likely reprice growth quality before revenue.