Why AI Profit Gains May Lag Behind Wall Street’s High Expectations

Wall Street is currently pricing in a massive surge in corporate profitability driven by artificial intelligence, but a growing warning suggests this timeline may be overly optimistic. Torsten Slok, chief economist at Apollo, argues that the actual realization of AI-driven margins in non-tech sectors could take much longer than current market projections suggest.

The Disconnect Between Tech Valuations and Real-World Adoption

The current valuation of many AI companies rests on a singular, massive assumption: that the S&P 493—the index excluding the "Magnificent Seven" tech giants—will see significant margin expansion through AI integration. While tech companies are already seeing the benefits of automation, the broader economy tells a different story.

According to Slok, there is currently no significant evidence that AI is boosting profit margins in industries outside of the core technology sector. Markets are pricing in rapid earnings growth, yet the actual cash flows required to justify these valuations may trail far behind. If the productivity revolution takes five years to materialize instead of five months, the industry could face a painful and significant repricing of AI stocks.

Structural Barriers in Regulated Industries

The delay in productivity gains is not merely a matter of slow adoption; it is a byproduct of the complex environments in which many global industries operate. In highly regulated sectors such as healthcare, banking, energy, pharmaceuticals, and manufacturing, the path to AI integration is fraught with hurdles.

Strict privacy requirements, rigorous regulatory compliance, and the necessity for massive process overhauls mean that these industries cannot pivot as quickly as a software startup. These institutional frictions act as a brake on the speed at which AI can be deployed to optimize workflows and reduce costs. Consequently, the "productivity bump" that investors are betting on may be delayed by years as these sectors navigate the legal and operational complexities of machine learning implementation.

The Measurement Problem in Knowledge Work

Even when AI successfully increases the efficiency of individual workers, a secondary challenge emerges: the difficulty of quantifying these gains. In the realm of knowledge work, productivity improvements are notoriously hard to measure with precision.

Unlike manufacturing, where output is easily tracked, the gains made by an employee using an LLM to draft reports or analyze data often lack clear metrics. Without concrete data, management teams struggle to justify large-scale capital expenditures or recognize the improvements on a formal balance sheet. Instead, these incremental efficiency gains are often "absorbed" into daily operations, failing to manifest as the visible margin expansion that Wall Street demands.

Key Takeaways

  • Market Overvaluation Risk: AI stock valuations are heavily predicated on margin growth in the S&P 493, which has yet to materialize in non-tech sectors.
  • Regulatory Friction: Industries like healthcare and banking face significant delays in AI adoption due to privacy mandates and complex process overhauls.
  • The Quantifiability Gap: Productivity gains in knowledge work are difficult to measure and report, making it hard for companies to translate AI usage into bottom-line growth.