Most large US companies have not yet seen financial benefits from AI - Apollo

The mass adoption of artificial intelligence has not yet provided noticeable profit growth for most large American companies. Despite significant investments in new technologies, the financial benefits of AI remain concentrated mainly among the largest technology corporations.

This conclusion was reached by Torsten Slok, chief economist and partner at the investment firm Apollo Global Management. His analysis was released on October 24, TechCrunch reports.

For the study, the economist divided the companies in the S&P 500 index into two groups. The first included the so-called "Magnificent Seven" - Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla, which play a leading role in the development and financing of artificial intelligence.

The second group Slok provisionally called S&P 493. It brings together the remaining 493 index companies, representing a broad spectrum of the American economy - from industry and commerce to financial services and the consumer sector.

The comparison revealed a significant difference in the financial results of the two groups. The largest technology corporations have growing revenue per employee and profitability. By contrast, the aggregate indicators of the rest of the companies remain practically unchanged.

To assess labor efficiency, the researchers used revenue per employee adjusted for inflation. According to Apollo's assessment, this indicator does not show noticeable improvement outside the group of largest technology companies.

A similar situation is observed with profitability. Despite the spread of artificial intelligence tools in the corporate sector, the overall profitability of companies from the S&P 493 group has not increased significantly.

Slok also analyzed the overall productivity of the American economy using data from the Federal Reserve Bank of San Francisco. He concluded that statistics do not yet show convincing signs of accelerating productivity due to AI adoption.

The lack of noticeable financial returns is forcing some companies to review their spending on artificial intelligence. According to the economist's observations, business is increasingly considering the payment for the use of AI models not as an investment in future growth, but as an additional item of operating expenses.

In particular, this concerns the payment for tokens - units of information processing by which AI service providers usually calculate the cost of using their models.

At the same time, technology startups that raise venture funding continue to increase spending on such services. Traditional corporations, however, do not always see the economic feasibility of using the most expensive and powerful models.

In Slok's opinion, the main question is whether companies will be able to turn the increasingly accessible AI tools into increased sales and profits.

The economist points out that business efficiency depends not only on the speed of information processing. Managerial decisions, customer trust, corporate culture, responsibility, and the ability to assess risks remain important.

At the same time, Apollo's conclusions do not mean that artificial intelligence brings no benefit to companies at all. The study shows the absence of a noticeable aggregate financial effect in a broad group of large corporations. Individual enterprises may benefit from automation, even if it is not yet reflected in overall profitability indicators.

Moreover, the growth in financial indicators of tech giants is not necessarily fully explained by the use of AI itself.

Assessments of the long-term prospects of the technology among economists differ. In particular, Goldman Sachs chief economist Jan Hatzius, as reported by Fortune, expects that investment in artificial intelligence will eventually provide increased productivity and profits across the entire economy.

Meanwhile, Goldman Sachs predicts that the contribution of large-scale AI investments to US economic growth may peak in late 2026 or early 2027, after which it will gradually decline.

Thus, one of the main tests for the artificial intelligence market will be the ability of the technology to provide measurable financial results outside of companies that directly earn from its development and infrastructure.

Sources: Apollo Global Management, TechCrunch, Fortune