Santoli: AI bull market needs AI trade to recover, with 45% S&P weight

Santoli frames the current pullback as a test of the AI-led bull market's leadership. He notes that in prior phases of the buildout, a key bottleneck gave investors a specific asset to buy: GPUs pointed to Nvidia from 2023 through 2024, cloud capacity and power and cooling made hyperscalers and AI-adjacent industrials work, and memory as the chokepoint sent Micron and Sandisk soaring. Now, as Nvidia's Jensen Huang has said, capital is the constraint, which Santoli says leaves no asset to buy and raises the cost of capital broadly as mega-caps go free-cash-flow-negative.

He also points to scrambled correlations: Treasury yields no longer move with oil but largely with AI-capex sentiment, and AI-stock setbacks mute inflation momentum and the equity wealth effect, meaning fewer Federal Reserve rate hikes are anticipated now versus a month ago. Amid this, Santoli says a lot of the known challenges may already be reflected in prices, with the S&P 500's forward P/E at 18.5 versus 23 a year ago and the Nasdaq-100 at 20x marking its slimmest premium to the S&P 500 in nine years. He notes sentiment toward AI stocks is in the dumps, funds' tech positioning has retreated below neutral, and the Nasdaq-100's relative-strength reading is exactly as oversold as at the March 30 and July 29 lows this year, so he advises staying alert for relief bounces that could become more consequential if the midterm elections pass in a way that lets traders unclench into the year's home stretch.