Torsten Slok, Chief Economist at Apollo Global Management, has published a study stating that the so-called "AI revolution" has brought measurable benefits to only a narrow circle of market participants. According to his data, the vast majority of companies in the S&P 500 index see virtually no financial results from the adoption of generative artificial intelligence and large language models, despite colossal financial investments and the hype surrounding the achievements of the AI industry. The macroeconomic impact of the technologies, the analyst argues, remains extremely limited: outside the tech segment, the American corporate sector has yet to receive a tangible financial return from artificial intelligence systems.
The Magnificent Seven vs. the S&P 493
Slok identifies a group of seven companies in the S&P 500—Nvidia, Tesla, Amazon, Microsoft, Google (Alphabet), Apple, and Meta—commonly referred to in business circles as the "Magnificent 7." According to his data, it is precisely these corporations whose inflation-adjusted revenue per employee shows steady growth, alongside improving profitability metrics. For the remaining 493 companies in the index, which Slok ironically designates as the "S&P 493," these indicators have remained at previous levels and show no dynamics that could be linked to the implementation of AI. Thus, the economic effect of the technologies is concentrated in a narrow segment rather than distributed across the entire economy.
From Investment to Cost: A Paradigm Shift
Slok pays special attention to the fact that he sees no signs of labor productivity growth driven by AI adoption. Lacking tangible returns, S&P 500 companies are beginning to rethink their approach to implementation spending: they increasingly view AI costs not as strategic investments capable of ensuring future business growth, but as expenses that reduce current margins. This is leading to budget cuts for AI projects in traditional corporations. Unlike startups that scale up AI spending with venture capital support, established organizations do not feel an urgent need for the latest models and are not willing to pay premium prices for them.
A "Powerful Calculator" Rather Than a Strategic Partner
Slok does not rule out that over time, the use of AI models will become a mass and standardized product, and the cost of using them will drop to a level that allows companies to increase revenue and profitability. However, in his assessment, achieving this goal may be significantly more difficult than the industry's optimistic narrative suggests. The analyst characterizes AI as a powerful "calculator," emphasizing that sustainable business is based on sound judgment, not just computation. "AI does not understand concepts such as responsibility, trust, corporate culture, unwritten rules, ethical compromises, or fairness. Yet these very things determine reputational risks and ultimately affect revenue growth and profitability," Slok writes in his study.
Who Will Decide AI's Fate: The Seven or the 493
Slok concludes that the fate of artificial intelligence as an economic tool will be determined not by the Magnificent Seven, but by the remaining 493 companies in the S&P 500 index. If the scope of AI application continues to be restricted to only a small fraction of the economy, its macroeconomic impact will remain negligible. Apollo, the asset management firm where the analyst works, controls around one trillion dollars in assets, giving his assessments additional weight. His findings echo recent forecasts by Gartner analysts, who suggested that by 2030, more than $3.5 trillion invested in AI infrastructure could be wiped out. At the same time, not all economists share this pessimism: a number of experts believe that in the long run, investments in AI will pay off through productivity growth and increased profits across the entire economy, and the current gap between the "seven" and the rest of the companies is a temporary phase of adaptation.
