Apollo: Outside Seven Big Tech Names, AI Spending Has Yet to Move the Needle

Torsten Slok, Apollo’s chief economist, published an analysis on October 24 (US time) weighing up what American companies have actually gotten out of their AI spending. Apollo found that revenue per employee and profit margins are rising at the seven major US tech companies, but that both measures are largely flat across the other 493 companies in the S&P 500. The analysis asks whether money spent on AI has translated into better results for corporate America as a whole.

Has AI spending lifted revenue and profits?

The comparison rests on two yardsticks: how much revenue each employee generates, and what share of revenue ends up as profit. The point is to go beyond claims that AI has made work easier and ask whether companies are actually making more money. On both measures, Apollo says, the seven major tech companies and everyone else are moving in different directions.

Apollo also looked at the efficiency of the broader US economy. On the gauge that shows how much more output you get from the same amount of labor and capital, the firm concludes, AI’s effect is not clearly visible either. In other words, corporate results and economy-wide efficiency were examined side by side.

The charts behind the analysis draw on data from Bloomberg and Macrobond, the Federal Reserve Bank of San Francisco, and the US Bureau of Labor Statistics. Apollo’s conclusion is that once you bundle many companies together, clear improvement is hard to find.

What is different from the February and August analyses

Apollo made a similar argument back in February, finding no sign of an AI effect in the profit margins of companies outside the tech sector. In August it reported that the margins of companies buying AI had yet to improve. The firm has, in short, checked repeatedly this year on whether the profitability of AI spenders is actually getting better.

This time it widened the range of indicators. Alongside profit margins, it presented revenue per employee, economy-wide productivity, and data on corporate spending on AI usage.

Does any of this apply to Korean companies?

The subjects of the analysis are US companies and US economic data. It is not a direct measurement of the returns on AI investment at Korean companies. Still, any company using AI in its day-to-day work could adopt revenue per employee and profit margin as items on its own performance checklist.

Two things are worth watching from here: whether revenue per employee and profit margins also start climbing at companies outside those seven major tech names, and how precisely AI’s share of any such change can be isolated.