Key Points

• Apollo has found that the profits brought by AI are mainly concentrated among technology giants.
• Revenues, number of employees, and profit margins remain relatively stable.
• Enterprises are increasingly questioning the return on investment of AI expenditures.

The AI boom in American business is widening the gap in financial performance: the largest tech companies are recording stronger revenue and profitability, while most other large American enterprises have not yet shown similar gains.

In a research report on October 24th, Apollo Global Management Chief Economist Torsten Slok found that among the “Big Seven of U.S. Stocks” ( Magnificent Seven ), both revenue per employee and profit margins had increased, while the remaining 493 companies in the S&P 500 remained relatively stable overall.

This discovery raises a question that companies that have invested in the AI tool cannot avoid: When will this so-called improvement in efficiency actually be reflected in the financial statements?

Apollo claims that AI's profits are concentrated among tech giants

The analysis by Apollo compared the financial performance of the “Big Seven” with that of the remaining companies in the S&P 500.

The first chart shows that, after adjusting for inflation, revenue per employee has increased among tech giants, while it has remained relatively unchanged in the broader group of companies.

The second comparison shows a similar pattern of differentiation in profit margins: the profit margins of the seven largest technology companies have significantly improved, while the overall progress of other companies has been limited.

Apollo also examined total factor productivity – an indicator that measures economic efficiency and takes into account labor and capital inputs – but found no signs of acceleration that could be clearly attributed to AI.

This pattern has intensified previous concerns from the outside world regarding the investment returns of AI, especially in light of companies continuously allocating their budgets to software subscriptions, automation, and computing infrastructure.

However, these results do not prove that AI has not created value for individual companies. Overall financial data may mask improvements in productivity, which in turn could be offset by wages, financing costs, or other expenditures.

Why are companies re-examining AI expenditures?

Apollo believes that AI expenditures are still concentrated in the hands of a relatively small number of enterprises, while mainstream companies have become more cautious when purchasing increasingly complex models.

Slok also points out that there are differences between companies supported by venture capital and mature enterprises: startups are increasing their expenditure on AI tokens, while large organizations are questioning whether high-end models are worth the cost.

This difference is also reflected in recent studies on corporate spending: although the AI practice continues, the growth rate of spending has slowed down.

A September analysis cited by Forbes shows that 69% of S&P 500 companies are actively deploying AI, but only 29% have reported quantitative results, and merely 2% continue to track a certain performance indicator.

These data indicate that widespread adoption does not necessarily translate into measurable financial returns.

At the same time, AI infrastructure investment continues to expand. Amazon's capital expenditure plan illustrates that companies may invest substantial funds as cloud demand accelerates, thereby putting pressure on free cash flow.

What will happen if the profits of AI remain concentrated?

This financial impact is not limited to companies that purchase AI software.

Semiconductor manufacturers, cloud service providers, and data center suppliers may generate revenue from infrastructure expenditures before their customers have demonstrated that their applications are profitable.

Coinpaper previously analyzed the expenditure flow of AI, including payments to chip manufacturers, utility companies, and construction firms.

However, the sustainability of such expenditures ultimately depends on whether customers can create sufficient economic value.

The recent decline in semiconductor stocks indicates how soon investors will begin to question the financial assumptions of the industry.

This centralization has also raised broader concerns about financial stability, as expensive projects often rely on borrowing and expectations of substantial future profits.