Why Upstart Stock Plunged 23% in July

Upstart’s Q2 revenue rose 42% YoY and net income nearly tripled to $16.5M, but the stock is still down 23% in July amid macro pressure on lenders tied to high interest rates and fears agentic AI could replicate SaaS-like workflows. Management is targeting 40% full-year revenue growth for 2026 (35% CAGR through 2028) and reported 50% higher originations, with secured auto/home originations up 218% YoY. Despite a post-earnings bounce, shares remain well below prior highs as investors weigh improving performance against underwriting/AI substitution risks.

Analysis

UPST is less a pure “AI winner” than a levered expression on credit-market normalization: when lender appetite improves, its take-rate model can translate volume into operating leverage far faster than traditional banks. The key near-term variable is not the headline rate path alone, but whether securitization markets and funding partners keep absorbing production at attractive spreads; that is the bridge between growth and durable GAAP profitability.

The market is probably overpricing the risk that agentic AI commoditizes the business. Generic agents can automate workflows, but they do not replicate a loan-performance dataset across vintages and borrower cohorts, which is the part that actually drives underwriting edge; the real competitive threat is from larger lenders/platforms with distribution and balance-sheet flexibility, not from SaaS substitution. If UPST keeps proving it can expand into secured auto/home while reducing its own capital intensity, smaller consumer-finance platforms without proprietary data are the likely losers.

This remains a credit-cycle trade first and an AI trade second. In the next 1-3 months, upside likely depends on stable delinquency data, tight ABS spreads, and continued originations momentum; over 6-18 months, the thesis breaks if growth decelerates toward market levels or if profitability remains intermittent. The contrarian view is that the stock may still be under-owned versus its earnings power if rate cuts or spread compression arrive, but it is equally vulnerable to a single macro shock because valuation expansion is ahead of balance-sheet proof.

Key Decisions for Investors

• Tactically long UPST on pullbacks, 3-6 month horizon, but size it as a high-beta credit-cyclical; risk/reward improves only if originations and funding spreads stay constructive.
• Use a call spread rather than outright stock if taking exposure: buy 3-6 month UPST call spreads to express upside from continued operating leverage while capping downside if credit data rolls over.
• Set a falsifier alert on sequential originations and ABS/funding costs; if originations growth drops below ~25-30% or financing spreads widen, reduce or exit because the earnings bridge breaks quickly.
• Pair trade for the medium term: long UPST vs. short a broad SaaS/AI-multiple basket proxy if the market keeps treating UPST as a generic AI software name rather than a credit data platform.
• No aggressive short until there is evidence of funding-partner retreat or delinquency deterioration; otherwise the stock can keep rerating on incremental proof of model durability.