AI Pays Off Almost Nowhere Outside Tech, An Economist Says. The Spending Keeps Exploding. What The Numbers Really Show

🔍 Read the full analysis: AI Pays Off Almost Nowhere Outside Tech, An Economist Says. The Spending Keeps Exploding. What The Numbers Really Show 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 artificial intelligence has produced no visible financial payoff across corporate America outside technology, even as major companies keep committing large sums to computing infrastructure. The claim describes aggregate trends, not a finding that every non-tech company has received no benefit: available company-wide measures cannot isolate AI’s effects from other economic forces.

Apollo’s analysis compares the Magnificent Seven technology companies with the other 493 companies in the S&P 500. In charts cited by Slok, revenue per worker and profit margins have risen for the seven, while the corresponding measures for the other 493 have remained broadly flat. Apollo also points to productivity measures from the San Francisco Federal Reserve, which it says show no clear AI-related boost to productivity growth.

The data do not establish that AI caused the difference between the groups. Revenue and margins can move for many reasons, and a flat average can conceal companies that gained from AI alongside those that did not. The note’s use of “zero” is Slok’s interpretation of broad results, rather than a direct measurement of AI’s effect on every business. Ramp spending data cited by Apollo also cover Ramp’s business customers, not the S&P 493 as a whole.

The spending picture is clearer in corporate filings. Meta reported $34.91 billion in capital expenditures in the third quarter and forecast $135 billion to $145 billion in capital spending for 2026. Its October 22 filing lists roughly $825 billion in future commitments, including leases and contractual obligations for data centers, energy, cloud capacity, servers and network equipment. The total includes some spending not tied to AI, and it represents commitments over time, not a single-year bill.

Meta’s third-quarter revenue was $64.60 billion, up 26% from a year earlier, while operating cash flow was $32.79 billion. With capital expenditures included, free cash flow was negative $2.12 billion, compared with positive $10.6 billion a year earlier. The company’s headcount was 66,775, down 15% year over year. These figures show the scale of investment and its effect on cash flow; they do not by themselves establish whether the spending will earn an adequate return.

The Cost of AI’s Buildout

The contrast between large, measurable infrastructure costs and uncertain returns matters to investors, workers and customers. If AI spending produces lasting productivity or revenue gains, expensive data centers and computing services could support new products and more efficient work. If those gains do not materialize, companies may face weaker cash flow, pressure on margins and a greater need to borrow or attract outside funding.

That issue is no longer confined to chip supply and electricity. The source material points to financing as a growing pressure point: SpaceX borrowed $40 billion to buy AI chips, while reports said the cost of insuring debt issued by Meta and cloud provider CoreWeave had risen. Those facts do not establish that either company cannot meet its obligations, but they show why lenders and markets are watching the relationship between investment, cash generation and future demand.

There is also evidence of gains in particular uses. Studies summarized in the source report found customer-support workers resolved 14% more issues per hour, with a 34% gain among novice workers. Retail experiments ranged from no sales increase to a 16.3% increase, depending on the workflow. These results suggest that the business case may depend on how AI is deployed, rather than on adoption alone.

What Apollo’s Measure Captures

Slok’s October 24 note, titled “No Signs of AI Payoff Outside the Tech Sector,” drew on five charts. Alongside comparisons of revenue per worker and profit margins, it cited productivity indicators, the concentration of AI-related spending and the funding profiles of heavy spenders. Apollo’s interpretation is that the spending and financial gains are concentrated among a small set of technology firms and venture-backed companies, rather than broadly distributed across established businesses.

According to Ramp data cited in the note, the top 10% of firms accounted for 99.5% of spending on model serving and inference, 99.0% of neocloud spending and 95.6% of spending on OpenAI and Anthropic. The source says 80.4% of the top 1% of AI spenders were venture-backed, compared with 13.1% of median spenders. These figures describe the businesses using Ramp and the categories it tracks; they should not be read as a complete census of US corporate AI spending.

Slok’s note followed a rapid expansion in plans for AI computing infrastructure. Meta’s commitments include leases that have not yet started, some extending through 2036, and non-cancelable contracts, with a portion due in 2027. Separately, projections cited in the source put combined 2026 capital spending by Amazon, Alphabet, Meta, Microsoft and Oracle at nearly $780 billion. That figure is a forecast, not a reported commitment or final spending total.

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

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

Returns Remain Hard to Isolate

How much AI itself has changed company performance remains unclear. Revenue, margins and productivity are affected by labor costs, interest rates, tariffs and other factors. Comparing technology companies with the rest of the S&P 500 cannot separate those influences or show what would have happened without AI.

The source also does not provide a full breakdown of Meta’s commitments by AI and non-AI use. Its filing includes infrastructure for several parts of the business, and the commitments stretch across multiple years. Nor do the cited workplace studies establish that the same results will occur in other companies, industries or tasks.

The source says that 83% of nearly 9,000 executives reported no productivity impact from AI so far, with the average estimated effect at 0.2% a year. It does not provide enough methodological detail here to assess how respondents defined AI use or productivity. The figures are best treated as survey results, not a definitive measure of economy-wide impact.

Evidence to Watch in Filings

The next useful signals will come from companies’ capital-spending updates, cash-flow reports and productivity disclosures. Investors will be watching whether AI-related revenue and measurable efficiency gains grow enough to support the spending now being planned. Meta’s $135 billion to $145 billion 2026 capital-spending outlook provides a near-term benchmark, though it is a forecast rather than a final result.

Further reporting is also needed on the financing conditions facing data-center builders and AI infrastructure providers, and on how much of their capacity is being used by customers willing to pay for it. Until those details and longer-term operating results are clearer, Slok’s conclusion remains a warning about the lack of broad aggregate evidence so far—not a final verdict on AI’s returns.

Key Questions

What does Slok mean by “zero financial payoff”?

Slok is referring to aggregate measures in Apollo’s analysis: revenue per worker and profit margins have risen for seven large technology companies but have been broadly flat for the other 493 S&P 500 companies. The figures do not prove that no individual non-tech company has benefited from AI.

Does Apollo’s analysis prove AI has not raised productivity?

No. It says the productivity measures Apollo reviewed show no clear AI-related boost so far. Economy-wide data can lag, and those measures cannot isolate AI from other changes in the economy.

What is included in Meta’s roughly $825 billion in commitments?

Meta’s filing lists leases that have not yet started and contractual commitments, mostly for items such as data centers, energy infrastructure, cloud capacity, servers and network equipment. The total covers obligations over time and includes some spending that is not AI-related.

Are there examples of AI producing measurable workplace gains?

Yes. The source reports a customer-support study in which workers resolved 14% more issues per hour, with a 34% gain among novice workers. Retail experiments reported outcomes ranging from no sales increase to a 16.3% increase, depending on the workflow.

What evidence would clarify whether AI spending is paying off?

Future company results showing sustained revenue growth, improved productivity or lower costs linked to AI would help. Updates on capital spending, cash flow, customer demand and the use of computing capacity will also matter.

Source: ThorstenMeyerAI.com