We found the AI factor (and it was there all along)

A once secondary statistical factor now explains more US market risk than the so-called market factor

US equity portfolios likely carry exposure to AI despite appearing diversified, whether its manager intended to or not. Our analysis finds that a statistical risk factor tied to the AI build-out, spanning chipmakers, power infrastructure, memory, and networking equipment, has grown from a minor influence into the dominant source of systematic risk.

Using the latest Axioma Short Horizon Statistical Model, we traced how this factor evolved from an ambiguous growth-versus-value axis in late 2022 into a clearly defined AI-hardware axis today. Since mid-June 2026, it has explained more Russell 1000 index variance than the 1^st principal component, traditionally assumed to be “the market factor”.

For portfolio managers, the implications are significant. Idiosyncratic-looking bets on AI-linked stocks are increasingly systematic according to the statistical model, and a fundamental risk model alone may understate the concentration and correlation building up inside seemingly diversified portfolios.

Read this report to discover:

• How an equal-weighted portfolio of AI-linked stocks reveals a persistent gap between statistical and fundamental risk forecasts.
• Why a factor that looked like a growth-versus-value trade in 2022 has concentrated into a clearly defined AI-hardware exposure.
• How Cohen's d and correlation analysis can turn an unlabeled statistical factor into something a portfolio manager can name and monitor.
• How the statistical model's higher risk forecasts compare with the fundamental model when measured against realized portfolio returns.