Can AI Spending Deliver Returns Beyond Tech?

🔍 Read the full analysis: Can AI Spending Deliver Returns Beyond Tech? on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Apollo chief economist Torsten Slok said in an October 24 research note that available aggregate data show no signs of a financial payoff from AI outside the technology sector, even as companies commit more money to computing infrastructure. The finding describes broad patterns, not a measured causal effect: it does not establish that every non-tech company has failed to benefit from AI.

Apollo’s comparison shows revenue per worker and profit margins rising for the Magnificent Seven technology companies while remaining broadly flat for the other 493 S&P 500 companies. Reading the charts cited in the source material, the seven were near $780,000 in inflation-adjusted revenue per employee in the third quarter of 2026, compared with about $525,000 for the other firms. The seven’s margins were around 30%, against roughly 10% for the rest. These differences do not, by themselves, show that AI caused the gap.

The note also points to productivity measures from the San Francisco Federal Reserve, which it says show no signs yet of AI boosting productivity growth. Separately, data from Ramp, a corporate spending platform with more than 85,000 U.S. business customers, indicate that spending on AI services is concentrated among a small group of companies. The top 10% of firms account for 99.5% of model-serving and inference spending in that dataset. Ramp’s figures describe its customers, however, and should not be read as a census of all U.S. companies.

The investment bill is visible in company disclosures. Meta reported $34.91 billion in third-quarter capital expenditures and negative free cash flow of $2.12 billion, compared with positive free cash flow of $10.6 billion a year earlier. Its 2026 capital-spending forecast is $135 billion to $145 billion. A filing lists roughly $355.73 billion in leases not yet started and $454.52 billion in non-cancelable contractual commitments, along with up to $14.72 billion in contingent cloud purchases. Those figures total about $825 billion, but the commitments extend over time and are not all exclusively for AI.

Where AI Returns Are Showing

The gap between large technology companies and the rest of the market matters because the financial case for AI spending depends on businesses outside the firms building and selling the technology also gaining enough to justify their costs. If companies pay for models, computing and implementation without raising output or revenue, those expenses can put pressure on margins. That is the concern Slok described when he said some mainstream companies are beginning to treat token spending as a cost rather than a growth source.

There are, however, signs of returns in particular tasks. The source material cites research finding 14% more customer-support issues resolved per hour, with a 34% increase for novice workers, and retail experiments reporting sales effects ranging from zero to 16.3%, depending on the workflow. Those results suggest AI can help in defined settings, but they do not establish broad gains across firms or industries. The practical question for employers is whether benefits persist after implementation costs and whether they can be measured in their own operations.

The scale of investment also makes the financing question relevant beyond technology firms. Meta’s spending plans and future commitments, alongside a reported $40 billion SpaceX borrowing for AI chips, show how much capital is being directed toward infrastructure. If expected demand or returns fall short, companies and lenders could face financial pressure. The source material reports that insurance costs on debt issued by Meta and CoreWeave rose, but the provided information does not establish that AI spending alone caused those movements.

What Apollo’s Data Compare

Apollo’s October 24 note, titled “No Signs of AI Payoff Outside the Tech Sector,” was covered by news outlets the following day. Its headline rests on comparisons between the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—and the other S&P 500 companies, as well as on productivity indicators and spending data. The conclusion is about what can be seen in broad measures at this point, rather than an audit of individual AI projects.

That distinction matters because industry averages can conceal variation. A flat result for 493 companies could include some firms that gained from AI and others that did not, with the effects canceling out. The figures also cannot isolate AI from other influences such as labor costs, interest rates or tariffs. Nor are the spending patterns among Ramp’s customers necessarily representative of the entire economy.

Slok has also raised concerns about the economics of computing investment. In an October 20 note, he called the possibility that the marginal cost of computing exceeds its benefit to users a central flaw in the AI investment case. He later characterized some venture-backed spending on AI tokens as “circularity.” Those are his assessments of the market, not settled findings that apply to every AI provider or customer.

“AI has delivered zero financial payoff to corporate America outside of tech.”

— Torsten Slok, Apollo chief economist, in the October 24 research note

What the Aggregate Figures Miss

The central uncertainty is whether the lack of a visible economy-wide productivity lift reflects limited returns, a delay before returns appear, or gains hidden by aggregate data. The cited productivity series do not establish that AI had no effect on any business. Likewise, Slok’s “zero” is a characterization of broad financial patterns, not a direct measurement of AI’s causal impact on all companies outside technology.

The source material does not provide a common accounting of how much companies have spent on AI systems, how much of their spending is infrastructure for other uses, or what returns those investments have generated. Meta’s commitments include leases and contracts extending into future years, and some spending is not solely AI-related. The exact timetable and cost of any future payments also differ by commitment.

Several other figures require caution. The cited estimate that major cloud companies could spend nearly $780 billion in 2026 and more than $1 trillion annually from 2027 through 2030 is a forecast, not a confirmed bill. The source material also reports that 83% of nearly 9,000 executives saw no productivity impact so far and that the average reported impact was 0.2% a year, but it does not supply enough methodological detail here to interpret those results as a definitive measure of AI’s economy-wide effect.

Evidence to Watch in 2027

The next test will be whether companies report sustained gains in productivity, revenue or margins as AI tools move from trials into everyday work. Investors and executives are likely to track capital spending, cash flow, debt costs and utilization alongside revenue growth, rather than treating infrastructure commitments as evidence of returns on their own.

Upcoming company filings and earnings reports may clarify how much of current investment is being used, how commitments are scheduled and whether AI-related services are generating revenue outside the technology sector. Independent productivity data will take longer to show whether any gains are broad enough to appear in economy-wide measures. Until then, the evidence supports a mixed picture: substantial spending, concentrated adoption and measurable benefits in some tasks, but no clear aggregate payoff beyond major tech firms so far.

Key Questions

Did Apollo prove that AI has produced no returns outside technology?

No. Apollo’s “zero payoff” describes its reading of aggregate trends, including revenue per worker, margins and productivity measures. Those comparisons do not isolate AI’s effect or rule out gains at individual companies.

What did Apollo find about non-tech companies?

Apollo’s charts show revenue per worker and profit margins rising for the Magnificent Seven while staying broadly flat for the other 493 S&P 500 companies. The figures do not establish why those patterns occurred.

Are companies already seeing measurable benefits from AI?

Some cited studies and experiments report gains in specific customer-support and retail workflows. The reported results vary by task, and they do not show that AI has raised productivity or sales across the economy as a whole.

What does Meta’s roughly $825 billion figure represent?

It combines leases not yet started, non-cancelable contractual commitments and possible contingent cloud purchases disclosed in Meta’s filing. The commitments extend over time and are not all exclusively for AI.

What evidence would show whether AI spending is paying off?

Company-level results showing lasting improvements in output, revenue or margins, considered alongside implementation costs and capital spending, would help. Broader productivity data may also clarify whether gains extend beyond a limited set of firms and tasks.

Source: ThorstenMeyerAI.com