Heavy Spending on AI, but Productivity Gains Remain Elusive

Torsten Slok, chief economist at “Apollo Global Management,” said on October 24 that spending on artificial intelligence is primarily boosting the revenue of technology providers. Revenue and profit per employee have risen among the group of companies known as the “Magnificent Seven,” while these metrics have remained unchanged at the other 493 companies. According to the analysis, despite major investments, there is no clear evidence that artificial intelligence is increasing overall productivity.

The cost of using artificial intelligence (AI) models is often calculated based on the number of tokens. A token is a small piece of text processed by the model. Costs rise as employees submit more queries and AI agents perform more complex tasks. According to Slok, ordinary companies are beginning to view tokens as an additional expense rather than a source of growth. If customers do not see improvements in productivity, revenue, or day-to-day expenses, providers’ sales growth may not last long.

A similar situation has emerged in India’s financial technology sector. Of those who participated in a “PwC India” survey, 84% said they had not yet seen measurable financial results. Nevertheless, 48% identified artificial intelligence built on a company’s own data and knowledge as a key strategic priority. Of the respondents, 30% believe the adoption of AI agents over the next 3-5 years will improve customer service. Another 22% expect it to reduce costs and increase productivity.

Trust and security remain among the main obstacles. Of the survey participants, 41% cited customer trust and acceptance of the technology as the biggest barrier to the widespread use of payments through AI agents. Another 33% identified fraud, security, and risk management as the main problem. AI agents can independently perform tasks consisting of several stages. However, access to company systems creates both additional costs and security risks. To reduce these risks, “Nvidia” has introduced a hardware and software solution called the “Open Agent Safety Platform” that monitors agents’ actions.

Rising token usage has forced companies such as “Microsoft,” “Meta,” and “Amazon” to limit employees’ use of artificial intelligence. “OpenAI” CEO Sam Altman has also described token prices as a serious problem for customers. At the same time, “Amazon,” “Google,” “Meta,” and “Microsoft” have allocated substantial funds to data centers, chips, and energy infrastructure. According to the analysis, artificial intelligence’s broader impact on the economy will depend on the actual financial results of the 493 companies outside the technology sector.