Morgan Stanley's latest analysis indicates that amid the rapid evolution of AI technology, investors must differentiate between infrastructure providers benefiting from AI development and service providers potentially vulnerable to AI disruption.
Recent quantitative research from Morgan Stanley reveals that within AI thematic investing, "service sector exposure" has become a decisive variable. Data shows a significant negative correlation between service sector exposure and year-to-date investment performance—the higher a theme's exposure to services, the greater the impact from AI disruption concerns. Among these, the "AI Adopters" theme, with a service sector exposure as high as 53%, has shown the weakest performance.
The report states that "AI Infrastructure" has a service sector exposure of only 14%. Not only does it boast the strongest performance year-to-date, but it also leads comprehensively across six quantitative assessment dimensions. This suggests that within AI investment themes, priority should be given to allocating to infrastructure areas that face minimal direct AI disruption risk and benefit from structural capital expenditure growth.
Market concerns about AI disruption have expanded from the initial software industry to broader service sectors, including financial advisory and brokerage services.
According to the report, as AI capabilities demonstrate non-linear acceleration, investors are reassessing the profit sustainability of service-oriented business models. This concern is directly reflected in the stock performance of related industries, making "service sector exposure" a key driver of thematic investment performance.
Morgan Stanley's quantitative team classified the 25 industry groups within the Global Industry Classification Standard (GICS) into "services" and "non-services," calculating the weight of services within the "AI and Technology Diffusion" theme and its four sub-theme portfolios.
The research found that although these themes all fall under the AI investment umbrella, their year-to-date performance shows a clear negative correlation with their level of service sector exposure.
Among AI-related themes, the "AI Adopters" sub-theme carries the highest service sector exposure at 53%, significantly higher than other themes. This theme focuses on companies adopting AI technology to improve operational models and efficiency. However, this high concentration in services has made it the worst-performing AI sub-theme.
The core issue lies in uncertainty. As companies accelerate AI adoption, concerns regarding competitive dynamics and pricing sustainability emerge—precisely the core anxieties currently being priced into the market. Stephen Byrd, Head of Thematic Research at Morgan Stanley, advises a more selective approach within the AI Adopters space, focusing on companies that possess pricing power.
In stark contrast, the report highlights that the "AI Infrastructure" theme has a service sector exposure of just 14%, the lowest among all AI themes, alongside the strongest year-to-date investment performance. This theme is more focused on areas like computing power, semiconductors, and supporting hardware, with less direct exposure to service sector disruption risks.
Crucially, AI infrastructure benefits from ongoing capital expenditure and structural demand for computing, semiconductors, and related hardware. This demand is not a short-term phenomenon but a long-term trend accompanying the development of AI technology.
Morgan Stanley employs six quantitative dimensions to assess thematic attractiveness: Information Ratio, Breadth of Earnings Estimate Revisions, Bottom-up Morgan Stanley Earnings Forecasts, Valuation Levels, Mutual Fund Positioning, and Factor Exposure. On this scorecard, the AI Infrastructure theme stands out within the AI space, demonstrating strong risk-adjusted performance in the near to medium term, robust earnings estimate revisions over the past three months, and favorable mutual fund positioning, particularly among top-performing funds.