Navigating the AI hype cycle: A case for active

The AI investment cycle is likely to shift from indiscriminate spending on data centers toward selective ownership of profitable adopters, disruptors, and infrastructure winners, making active management more valuable.

• Semiconductor and data-center ETFs have outperformed broad equities but fell 10.2% through October 16, while semiconductor and equipment stocks dropped more than 12% in July, highlighting their cyclicality and volatility.
• Investors directed nearly $82 billion into technology stocks in the 12 months through August 2026, more than three times the inflows into the next-leading sector, industrials.
• Value indexes may provide exposure to AI adopters, but Amazon and Microsoft are already two of the three largest holdings of the Russell 1000 Value Index because of their billions of dollars of AI data-center spending.
• The article argues that AI-driven margin gains will depend on management quality, data infrastructure, and competitive advantages; in highly competitive industries, shared cost savings may instead be passed to customers through lower prices.
• Hyperscaler valuations are around 20 to 25 times expected earnings, but potential overcapacity from trillions of dollars of data-center investment could produce disappointing aggregate returns and still leave important winners and losers.

Takeaway: Investors should favor active managers able to distinguish durable returns on investment, strong balance sheets, efficient products, adaptable infrastructure, and smaller AI-enabled businesses from shortage-driven or hype-driven beneficiaries, while recognizing that active fees and underperformance remain risks.