Affirm's AI Upgrade Strengthens Underwriting but Valuation Risks Persist

What happened

Affirm announced an upgraded AI model trained on 14 years of data that expands credit approvals and improves early loan performance, according to a Zacks report. This development aligns with the company's existing data-driven underwriting moat highlighted in the DeepValue analysis, which acknowledges the strength of its risk models. However, the broader assessment remains cautious: the stock trades at over 100x earnings and more than 200% above a DCF-based intrinsic value. The report notes high leverage (Net Debt/EBITDA ~9x) and reliance on wholesale funding in a sector with elevated delinquencies. While the AI upgrade may incrementally improve credit outcomes, it does not address the core risks of valuation and balance-sheet fragility.

Implication

Investors should monitor whether the AI upgrade leads to measurable improvements in delinquency and charge-off rates relative to peers, as this could strengthen the company's competitive position. However, even if underwriting improves, the stock's current price embeds optimistic assumptions that leave little margin of safety. The reliance on secured funding and thin equity buffer makes the equity vulnerable to credit or funding shocks. Any deterioration in sector delinquency trends or regulatory headwinds could trigger a significant repricing. Until the valuation reverts to a level that offers adequate compensation for risk, existing shareholders should consider trimming exposure and prospective buyers should await a more attractive entry point.

Thesis delta

The thesis shifts marginally: the AI upgrade reinforces the durability of Affirm's data advantage, which was already recognized in the moat assessment. However, the core thesis of overvaluation and high risk remains unchanged; the news does not alter the POTENTIAL SELL judgment. If the model demonstrably lowers losses and supports higher approval rates, it could improve long-term economics, but that would need to be reflected in future financials, not just headlines.