1. AI leadership is shifting from a "concentrated enabler" to a "barbell" structure. Morgan Stanley's Stephen Byrd mapped approximately 3,600 stocks to their AI exposure and importance, believing that semiconductors and infrastructure have outperformed the S&P 500 by about 500-700 points since the start of the cycle; consensus predicts that high-importance AI adopters will achieve a 460bps EBIT margin expansion in 2025-26. The next step should be to expand to early software enablers and AI adopters.
2. Fund flows validate the diffusion of the AI theme. Funds continue to buy proxy AI, semiconductors, and global memory, further increasing their exposure to less crowded AI sectors. Meta Muse's top ranking on the App Store has fueled proxy commerce, with MSXXCLAW (proxy AI beneficiary basket) outperforming the broad AI Capex basket by 10% in 10 days; the inverse MSXXARSK (proxy risk basket) includes CHTR, CMCSA, and Siri, betting on the impact of proxy services on subscriptions, advertising, and aggregation businesses. (Morgan Stanley report, September 26)
