Artificial intelligence has been hailed as a revolutionary force set to redefine industries and supercharge economic growth. But a recent analysis from a top economist offers a dose of reality, suggesting the promised payoffs are not yet materialising.
A Reality Check on the AI Boom
A widely circulated note from Torsten Sløk, chief economist at investment giant Apollo Global Management, has sent a ripple of caution through the business world. His analysis concludes that despite colossal investment and hype, artificial intelligence has so far delivered 'zero financial payoff' to companies outside of the major technology players. Sløk's argument is grounded in data, comparing the performance of the 'Magnificent Seven' tech stocks (like Nvidia and Microsoft) with the remaining 493 companies in the S&P 500 index. The findings are stark: while the tech behemoths have seen revenue per employee and profit margins soar since 2022, these metrics have remained almost entirely flat for everyone else. This suggests the AI boom is, for now, a concentrated phenomenon, with the profits flowing to the companies creating and selling the technology, not necessarily to the ones using it.
Following the Data, Not the Hype
Sløk's analysis points to several key indicators. Revenue per employee for the top tech firms has jumped from around $550,000 to nearly $800,000 in recent years, while it has hovered at the $550,000 mark for the rest of the S&P 500 companies. A similar story unfolds with profit margins, which have surged for the Magnificent Seven but stagnated at about 10% for others. Most critically, Sløk finds no clear evidence that AI has started to boost overall economic productivity, a key promise of the technology's proponents. This leads to his conclusion that 'mainstream corporate America is pulling back' and beginning to view AI spending less as a growth engine and more as a cost centre that is squeezing their profits. The data suggests that for many businesses, the return on their significant AI investments has been difficult to find on the balance sheet.
The Productivity Paradox
The gap between a technology's potential and its measured economic impact is a well-known phenomenon often called the 'productivity paradox'. History shows that even transformative general-purpose technologies like electricity and the internet took years, sometimes decades, to translate into broad productivity gains. Companies don't just plug in a new tool and see immediate results; they must also reinvent workflows, retrain staff, and overhaul business processes to fully leverage the new capability. This implementation lag is likely a major factor in the current situation. Some studies even show an initial dip in productivity as companies struggle to adapt. Furthermore, Sløk argues that AI fundamentally 'does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness', all of which are crucial for business success and drive long-term revenue and profits.
Is the Spending Bubble at Risk?
The staggering amount of capital being poured into AI infrastructure adds another layer of tension. Concerns are growing over debt-funded investments in data centres and chips, with some analysts warning that over $3.5 trillion in spending could be at risk by 2030 if the promised revenues don't materialise. The high costs are already forcing some companies to limit their use of the most advanced AI models. This dynamic creates a precarious situation: if non-tech companies continue to see AI as a cost without a clear benefit, they could slash their budgets. This would, in turn, threaten the demand for the very infrastructure that firms like Apollo itself are financing, creating a potentially risky circular logic. Even as Apollo's CEO expresses bullishness on financing AI infrastructure, the skepticism from their own chief economist highlights a growing doubt on Wall Street about when, or if, the widespread payoff will arrive.
