Executive Summary 

• Market interest and investment in AI has skyrocketed, with infrastructure companies (i.e., those enabling AI) reaping the immediate benefits.
• Innovation cycles often become over-hyped, with stock prices anticipating higher earnings that can take years to materialize​.
• The required earnings growth to justify current valuations of AI related companies appears lofty, but is not yet at extremes seen during the internet bubble.
• Investors should exercise caution around hot technology trends, as the price one pays for those investments is a key determinant of future returns.​

Shown in the left panel is U.S. manufacturing construction put in place for computer, electronic or electrical manufacturing in billions of U.S. dollars on a seasonally-adjusted basis. Shown in the right panel are annualized total returns since 12/31/2022 for various cohorts of the S&P 500 index split by AI infrastructure, Software and Rest of S&P 500. AI Infrastructure is represented by semiconductor firms, data center real estate investment trusts and hardware & equipment companies, Software is represented by software companies and Rest of S&P 500 is represented by the remaining companies not included in the prior cohorts. The S&P 500 is a market capitalization weighted indices of U.S. large cap stocks. Past performance may not be indicative of future results. One cannot invest directly in an index.

• The advent of generative AI has sparked a massive surge in tech-related investment, particularly in companies that provide the infrastructure necessary for AI development.​
• These AI infrastructure companies—encompassing semiconductor firms, data center REITs and hardware & equipment providers—have been the immediate beneficiaries of this wave.​

Innovation cycles often become over-hyped, anticipating higher earnings that can take years to materialize

Data shown are the earnings and market capitalization of the S&P 500 Index’s information technology sector. S&P 500 Information Technology Earnings is an estimate of market value based on net income capitalized at a rate of 20 times earnings. The S&P 500 and its sector-specific sub-indices are market capitalization weighted indices of U.S. large cap stocks. Past performance may not be indicative of future results. One cannot invest directly in an index. 

• Innovation cycles often become over-hyped, with stock prices anticipating higher earnings that can take years to materialize​.
• Earnings during the tech bubble fell short and even declined after the bubble popped, leading to a decade-long gap before catching up to lofty market expectations.

AI valuations build in high expectations for earnings, similar to but not quite as high as a prior boom

Data shown in the left panel are the earnings and market capitalization of the S&P 500 Index’s Information Technology sector. S&P 500 Information Technology Earnings is an estimate of market value based on net income capitalized at a rate of 20 times earnings. Data shown in the right panel are the earning and market capitalization of AI Infrastructure companies, which are represented by semiconductor firms, data center real estate investment trusts and hardware & equipment companies in the S&P 500 Index. The S&P 500 and its sector-specific sub-indices are market capitalization weighted indices of U.S. large cap stocks. Price-to-Earnings (P/E) ratio is calculated as market value over net income. Required EPS Growth represents the growth in EPS (earnings per share) needed over the next 5 years to generate a return of 9% annualized from the starting P/E ratio and assuming a P/E ratio of 20x at the end of the 5-year period. One cannot invest directly in an index.

• The required earnings growth to justify current valuations of AI related companies appears lofty, but is not yet at extremes seen during the internet bubble.​
• Investors should exercise caution around hot technology trends, as the price one pays for those investments is a key determinant of future returns.​