AI has delivered ‘zero financial payoff’ outside tech, Apollo economist says

Artificial intelligence has delivered “zero financial payoff” to corporate America outside technology, Apollo Global Management chief economist Torsten Slok said in an October 24 note. His assessment challenges the expectation that heavy investment in AI is already producing widespread improvements in business performance.

Writing in Apollo’s Daily Spark, Slok contrasted the Magnificent Seven with the remaining 493 companies in the S&P 500. Revenue per worker and profit margins are rising for the former but remain flat for the latter, according to his analysis.

Slok also said productivity figures show no clear AI boost. He argued that mainstream businesses are beginning to treat spending on AI tokens as a cost burden rather than an engine of growth, while venture-backed companies continue increasing their consumption.

The distinction matters because selling AI and profiting from its use are different economic propositions. Strong demand for a technology can benefit its suppliers before customers demonstrate that their purchases generate enough additional revenue or savings to cover the expense.

For a business buying AI services, the relevant comparison is not simply whether employees can complete a task faster. The financial test is whether that improvement ultimately changes costs, output or sales enough to improve results after the expense of using the technology is included.

There are important limits to the headline conclusion. Comparing broad groups of companies does not isolate AI’s contribution to their performance, and flat aggregate margins do not establish that every company or project has earned nothing. Benefits at some businesses could be offset by unrelated pressures elsewhere.

Likewise, stronger results among the largest technology companies do not, on their own, measure how much improvement came from AI. Slok’s comparison highlights a divergence in financial outcomes; it should not be read as a controlled test establishing the cause of that divergence.

Slok’s outlook also rests on a distinction between cheaper computational capabilities and effective business judgment. He argued that context, accountability and trust remain central to commercial decisions, complicating the prospect that increasingly inexpensive models will automatically translate into better financial performance.

That leaves adoption and returns as separate questions for companies assessing their AI budgets. A tool can be useful without materially changing company-wide earnings, while an initially expensive deployment could require more time before its financial contribution becomes visible. Neither possibility settles the aggregate debate today.

The next test, in Slok’s framing, is whether companies beyond the technology leaders can turn cheaper AI into faster revenue growth and higher margins. His conclusion is that the broader corporate sector, rather than the leading technology suppliers alone, will determine whether the investment wave delivers a widespread economic payoff.