Scott Rubner, head of equity and equity derivatives strategy at Citadel Securities, identified the upcoming US midterm elections as a clearing event for market risk. Speaking to Bloomberg's Dani Burger, Rubner argued that the current period marks the point when AI stocks can outperform again, provided investors wait for the political uncertainty to resolve.
Rubner described a sharp divergence between institutional and retail behavior during recent volatility. While institutional clients have deleveraged, often due to rules-based unwinding triggered by price or volatility levels, retail investors have continued to buy aggressively. He noted that retail clients purchased semiconductors, neo clouds, and memory stocks at the largest rate in 16 weeks during the previous week. This divergence contributed to a significant unwinding of momentum factors, which were up 70% between January 1 and June 22 before turning negative for the year.
Despite the concentrated pain in AI-adjacent names, Rubner characterized the broader market movement as a rotation rather than a flush. He cited an S&P 500 drawdown of 20% from highs for 38% of its constituents, while broad indices remained only 150 bips below their peaks. He emphasized that such a pattern, where a large portion of the index drops significantly while the overall index stays near highs, had not been previously observed. On average, an 80% down day in New York occurs about 21 times a year, but none had occurred in over a year.
Rubner pointed to earnings data as evidence that the average company is benefiting from AI productivity and revenue growth, distinct from those labeled as AI companies. He highlighted that Q3 earnings expectations show 27% year-on-year growth, which he described as the third highest since 2010. In Q2, nine out of eleven S&P 500 sectors reported double-digit revenue gains. He expressed confidence that this trend would continue into 2027, noting that peak earnings for that year had not yet been reached.
Positioning in tech has shifted dramatically, Rubner stated. Net exposure in tech is in the fifth percentile over the past 35 years. The broad S&P 52-week forward price-to-earnings ratio has fallen to 18 times from 23 times earlier in the year, representing a 20% derating. In information technology specifically, valuations are at their lowest level since 2022, below levels seen in March and April.
Rubner said client feedback indicates a willingness to buy higher but not into major political events. He anticipated a rotation back into unloved themes such as cyclicals, transports, and home builders if interest rates reverse. He noted that retail investors are increasingly allocating to fixed income, specifically 30-year inflation-adjusted notes collecting 3% above inflation. This shift reflects a move toward sophisticated portfolio allocation driven by AI tools, resembling institutional investor behavior rather than previous low-quality thematic trades.
He added that passive ETFs have seen record inflows year-to-date, which is 60% higher than last year's record. With households owning 53% of the US stock market compared to hedge funds' 2%, Rubner suggested that agentic AI is changing market structure by enabling daily behavioral changes and monthly rebalancing similar to target date funds.