Corporate America has poured billions into artificial intelligence over the last few years, yet the financial returns remain concentrated in a narrow slice of the market. According to a note released Saturday by Apollo Global Management, the broader business world has seen no tangible benefit from these expenditures.
Torsten Slok, chief economist at Apollo, wrote that despite widespread hype and massive capital deployment, the technology has delivered zero financial payoff to corporate America outside of the technology sector. The data suggests a sharp divergence in performance between the industry leaders driving the AI boom and the rest of the S&P 500.
Revenue Divergence Between Tech Giants and Mainstream Firms
Slok highlighted a stark contrast in productivity metrics. Revenue per worker is rising for the “Magnificent 7” companies, a group comprising Amazon, Apple, Google, Meta, Microsoft, Nvidia, and Tesla. In contrast, that same metric has remained flat for the remaining 493 companies in the S&P 500 index.
This disparity indicates that while tech giants are leveraging AI to boost output and earnings, other industries have not yet translated their investments into measurable financial gains. The note suggests that mainstream corporate America is beginning to reassess its approach to artificial intelligence.
Companies are increasingly viewing token spending not as a driver of growth, but as a cost center that weighs on profit margins. This shift in perspective is likely to influence how businesses allocate capital in the coming quarters, particularly as earnings season begins and investors scrutinize the return on AI expenditures.
Investor Skepticism and Meta’s Heavy Spending
Bloomberg News flagged Slok’s comments on Sunday, noting that investors are closely monitoring which companies can justify their AI outlays. Meta Platforms recently reported its third-quarter numbers, revealing $34.9 billion in capital expenditures while posting a negative free cash flow of $2.1 billion.
Despite the financial strain, Meta CEO Mark Zuckerberg announced plans to increase AI-related capital spending. During an earnings call, Zuckerberg highlighted the success of the company’s Muse AI model, which has been adopted by more than 10 million people in the U.S. and Canada within six weeks. Approximately a quarter of those users engage with the model daily.
However, the market reaction was negative, with Meta’s shares declining following the announcement. The report suggests that Zuckerberg is betting on Muse to revive Meta’s AI ambitions and generate sufficient revenue to justify the heavy spending. Whether the model can deliver the necessary financial returns remains unproven.
Adoption Rates vs. Actual Autonomy
While adoption of AI tools has surged, actual operational autonomy remains limited. Research by PYMNTS Intelligence shows that in 2025, only 7% of executives were using or piloting AI agents, with 67% refusing to grant them access to company systems.
By 2026, 89% of executives were using agents in at least one part of the business, and only 2% denied access. Yet, most companies have not allowed these agents to operate independently. Seventy-eight percent of firms let agents read company systems but not act on them. Another 18% allow agents to carry out tasks only with human approval. Just 2% permit agents to run routine work on their own.
This cautious approach reflects broader investor wariness. The collapse of an initial public offering by Australian data center company Firmus Grid, where investors rejected a $30 billion valuation, underscores the pressure on AI-related valuations. Data center developer Nscale is also facing investor protest over its proposed listing value.
Slok’s analysis underscores the current reality for most businesses: AI is a significant expense with limited immediate financial reward outside the technology sector. As companies face margin pressures, the focus is shifting from aggressive adoption to proving concrete returns on investment.
Source: PYMNTS
