AI payoff is zero outside tech, Apollo's Torsten Slok says

Key takeaways

• Apollo chief economist Torsten Slok wrote on October 24 that AI has delivered "zero financial payoff to corporate America outside of tech."
• His charts show revenue per worker and profit margins rising for the Magnificent Seven but flat for the other 493 companies in the S&P 500.
• Apollo data reported by Fortune put Magnificent Seven margins near 25% in early 2026, up from about 15% in 2023, against roughly 10% for the S&P 493.
• Slok's argument is that the AI boom has so far lifted the margins of AI sellers (chips, cloud), not AI buyers.
• The note lands days before Microsoft, Alphabet and Amazon report earnings, when AI capital spending is the market's main question.

The AI payoff has yet to reach companies outside tech, Apollo Global Management chief economist Torsten Slok wrote on October 24, finding "zero financial payoff to corporate America outside of tech." His charts show margins and revenue per worker flat at the S&P 493.

The argument matters for the S&P 500 because much of the index's weight sits in a few AI-linked stocks, and Slok has warned that a delayed payoff would hit the wider market, not just tech.

What Torsten Slok's AI payoff note says

The note, titled "No Signs of AI Payoff Outside the Tech Sector" and dated October 24, 2026, is part of Apollo's Daily Spark series. Slok divides the S&P 500 into two groups: the Magnificent Seven (Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla) and the remaining companies, which he calls the "S&P 493."

Five charts make up the case:

• Revenue per worker, adjusted for inflation to the fourth quarter of 2019, is rising for tech and flat for the rest of the index.
• Profit margins are going up for the Magnificent Seven and show no upward trend for the S&P 493.
• Productivity growth, measured by the San Francisco Fed's utilization-adjusted total factor productivity series, shows no AI boost so far.
• AI spending is highly concentrated in a small number of firms, based on data from fintech company Ramp.
• Token spending is circular: venture-backed companies are spending more on AI tokens, while mainstream firms are holding back.

Slok writes that mainstream companies are treating token spending as a cost that weighs on margins and are unwilling to pay for the latest models. He also argues that the technology lacks business judgment, writing that AI "does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness." His conclusion is that "the AI story will be decided in the S&P 493, not the Mag 7."

TechCrunch's report, written by Anthony Ha, quotes the same "zero financial payoff" line along with another from the note: "AI may be a great calculator, but in business, intelligence is mostly judgment."

What the margin data shows

The October note gives its evidence in charts and does not list the values in its text. Earlier Apollo work, reported with numbers, fills in the gap. In July, Fortune reported that Slok's Bloomberg and Macrobond data showed Magnificent Seven profit margins climbing from about 15% to 25% between the first quarter of 2023 and the first quarter of 2026. Over the same period the S&P 493 stayed near 10%, and the broader Bloomberg 500 index held at about 12%.

Source: Apollo Chief Economist using Bloomberg and Macrobond data, as reported by Fortune on July 6, 2026. Figures are approximate chart readings.

A sector breakdown in Apollo's August 9 note, "The Buyers of AI Are Still Waiting for the Payoff," shows the same pattern. Margins rose in information technology and communication services. Elsewhere, health care margins have roughly halved since 2015, consumer staples sit near 6%, consumer discretionary near 8%, and energy and materials have given back most of their 2022-23 gains. Slok described the few improvements as looking like an ordinary cyclical recovery rather than a technology-driven step change.

Why AI sellers are winning and AI buyers are waiting

The August note put the thesis in one line: the AI capex boom is "so far only showing up in the sellers' margins, not the buyers'." Chipmakers, server makers and cloud providers book revenue the moment a customer signs a contract or buys capacity, while the buyer still has to turn that spending into higher sales or lower costs.

The cost side is also proving harder to manage than expected. Apollo's October 14 note, "AI Spending: Waiting on ROI, Watching the Bill," cited a Mavvrik and Benchmarkit survey of 396 enterprises carried out in April and May 2026:

Source: Mavvrik and Benchmarkit, 2026 State of AI Cost Governance Report, cited by Apollo on October 14, 2026.

"We are still waiting for the ROI from AI investments," Slok wrote in that note. He has made the same point since at least February, when an Apollo chart showed consensus earnings expectations rising for the Magnificent Seven but "basically unchanged" for the S&P 493.

What it means for the AI capex boom

Spending is still accelerating on the seller side. Alphabet raised its 2026 capital expenditure guidance to between $195 billion and $205 billion in July, from $180 billion to $190 billion, according to its second-quarter earnings call as reported by W.Media. Microsoft, Alphabet and Amazon report again on October 28 and 29 (see our Big Tech earnings dates and capex tracker).

Slok's concern is the gap between that spending and the payoff. In a July 9 note, "A Slower AI Payoff Would Be Everyone's Problem," he wrote that consensus expects hyperscaler free cash flow to "more than double over the coming years." He listed three risks if returns arrive later than that:

1. Cash flow and earnings disappoint while capital spending and depreciation stay on schedule.
2. A sell-off in the Magnificent Seven spreads to chips, power, data centers and the wider S&P 500, because those stocks make up so much of the index.
3. Hyperscalers borrow more, raising leverage and the risk of credit-rating downgrades.

He also pointed to two forces that could slow the payoff: falling token prices and the growing use of cheaper Chinese models. The debt-funded side of the buildout, including GPU cloud providers, is covered in our explainer on what a neocloud is and how it makes money. For the hardware side of the seller story, see our report on Super Micro's preliminary revenue beat on AI servers.

What the data does not show

Slok's charts are aggregate measures, so they cannot rule out gains at individual companies or in specific functions such as customer service or software development. Profit margins also move with interest rates, wages and the business cycle, which makes it hard to isolate the effect of any one technology. Slok's argument is mainly about timing. "The key issue is the length of the ROI runway outside the tech sector," he wrote in July, as quoted by Fortune, warning that expectations running ahead of actual earnings could lead to a "painful repricing."

Readers should also note the source. Apollo is a large alternative asset manager with its own investment views, and the notes carry the usual caveat that they reflect the author's opinion as of the date shown. TechCrunch also pointed out that Apollo once owned the publication.

The next test comes this week. If big tech raises spending again without clearer evidence that customers outside tech are earning a return, the question Slok keeps raising will get louder. More coverage is on our Apollo Global Management topic page.

This article is for information only and is not investment advice.