Apollo's Torsten Slok sees no AI payoff for corporate America outside the tech sector

Apollo chief economist Torsten Slok says artificial intelligence has so far paid off only inside the tech sector [1]. In a Daily Spark note dated 24 October 2026, he wrote that revenue per worker and profit margins are rising for the Mag 7 but flat for the other 493 companies in the S&P 500 [1].

What did Torsten Slok find?

Torsten Slok wrote that AI has delivered zero financial payoff to corporate America outside of tech. He made the claim despite what he called massive investments and widespread hype around the technology [1].

Slok said revenue per worker and profit margins are rising for the Mag 7 but flat for the S&P 493. He added that there are still no signs of AI boosting productivity growth [1]. The note backs each point with a chart; it does not state the underlying values in its text [1].

Why does token spending matter for the argument?

Slok argued that mainstream corporate America is starting to view token spending as a cost center. In his words, it is seen not as a source of growth but as a cost that weighs on margins, because it has had no impact on the bottom line [1].

He contrasted that with venture-backed companies, which he said are spending more and more on tokens [1]. According to Slok, mainstream companies do not need the latest model and are unwilling to pay for it. As long as agentic AI and token spending stay confined to a small part of the economy, he expects the macroeconomic effects of AI to be very limited [1].

Two of the note's charts draw on Ramp data, including the Ramp AI Index [1]. Ramp's own index, built from transactions of more than 85,000 firms on its platform, shows how uneven spending is: among firms paying for AI in September 2026, the median firm spent 15.70 dollars per employee per month, while the top 1 percent spent a median of 9,567.53 dollars [2].

What would change Slok's view?

Slok framed the open question as whether mainstream corporate America can turn increasingly cheap intelligence into higher revenue growth and profit margins, even as models become commoditized [1]. He wrote that this may be harder than it sounds.

He argued that AI does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness, which he said shape reputational risk and ultimately revenue and margins [1].

Our read

Opinion: Slok is measuring the payoff where it should eventually show up, in revenue and margins outside tech, and his flat lines are a fair warning against reading AI capex as proof of AI returns. But flat lines can also be what an early adoption curve looks like, and Ramp's own numbers show a tiny group of heavy spenders pulling away from everyone else. The S&P 493 test is the right one; the verdict is not in yet.

What we could not verify

• Exact chart values: Slok's note publishes its evidence as five charts, and this article uses only the conclusions he states in the text, not figures read off the charts.
• How Slok defines the Mag 7 and S&P 493 groupings, the time window of each chart and how revenue per worker is calculated beyond the note that values are inflation adjusted to 2019 Q4.
• Whether corporate America is in fact pulling back on token spending; the note states this as Slok's assessment without a separate survey or company data in the text.
• Whether the Ramp AI Index, based on firms using Ramp, is representative of all US businesses.