AI’s Economic Impact: Spending Up, Payoffs Limited Beyond Tech

🔍 Read the full analysis: AI’s Economic Impact: Spending Up, Payoffs Limited Beyond Tech on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get smart everyday buys 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 Global Management chief economist Torsten Slok says recent company and productivity data show no visible aggregate financial payoff from AI outside the technology sector, even as businesses and infrastructure providers commit large sums to computing capacity. His October 24 analysis contrasts gains among the seven largest technology companies with flat results across the other 493 S&P 500 firms, but it does not establish that AI has delivered no benefits to individual businesses.

Apollo’s charts track inflation-adjusted revenue per worker and profit margins. The seven companies Apollo groups as the “Magnificent 7” — Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla — show rising revenue per employee and margins near 30%. The other 493 S&P 500 companies remain roughly flat on revenue per worker and around 10% on margins, according to Apollo’s reading of the data. Those differences describe company performance; they do not isolate AI as the cause.

The analysis also points to productivity indicators from the San Francisco Federal Reserve, which Apollo says show no signs of an AI-driven productivity boost so far. Separately, data from Ramp, a corporate spending platform, show AI-related spending concentrated among a small group of firms: the top 10% of its business customers account for nearly all reported spending on model inference and neocloud services. That is a finding about Ramp’s customers, not a representative measure of every US company.

Meanwhile, reported commitments show the scale of the infrastructure buildout. Meta’s October 22 filing lists approximately $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. Together, the figures amount to roughly $825 billion, though the commitments extend across years and include some activity not tied to AI. Meta separately forecast $135 billion to $145 billion in capital spending for 2026.

Big AI Bills, Uneven Business Returns

The divide between measurable spending and unclear returns matters for companies deciding whether to expand AI use, and for investors and lenders financing the infrastructure behind it. If the technology does not improve revenue, costs or output at a broad range of businesses, large commitments may take longer to earn back. That is a risk to evaluate, not a conclusion that the spending will fail.

There is evidence of gains in specific tasks. Research cited in the source material found customer-support workers resolved 14% more issues per hour, with a 34% gain for novice workers. Retail experiments reported sales changes ranging from no improvement to a 16.3% increase, depending on the workflow. These results suggest outcomes can vary with how AI is deployed; they do not show that the same gains apply across industries or persist at scale.

Other cited survey data indicate that most executives have not yet seen a productivity effect: 83% of nearly 9,000 executives reported none so far, while the reported average productivity gain was 0.2% a year. The survey’s results reflect respondents’ reports and its stated measurement period, rather than a direct estimate of AI’s total economic impact. Taken together, the figures frame a gap between investment and broadly visible returns.

What Apollo’s Data Actually Track

Apollo published its “No Signs of AI Payoff Outside the Tech Sector” note on October 24. Its central comparison is between the seven large technology companies and the rest of the S&P 500. Slok’s phrase “zero financial payoff” is an interpretation of those aggregate patterns, not a direct accounting of the effect of AI on every company.

That distinction matters because an average can conceal companies that are gaining and others that are not. The charts also cannot, on their own, distinguish AI’s effects from other influences such as labor costs, interest rates or tariffs. Ramp’s spending figures cover its own customer base, and the analysis does not test specific workplace deployments. The strongest supported description is that a broad payoff is not yet visible in the cited aggregates.

The investment side is moving quickly. Meta’s third-quarter results showed $34.91 billion in capital expenditures and $32.79 billion in operating cash flow, producing negative free cash flow of $2.12 billion for the quarter, compared with positive $10.6 billion a year earlier. Meta reported revenue of $64.60 billion, up 26% year over year, and headcount of 66,775, down 15% from a year earlier. These results provide financial context, but they do not attribute the quarter’s performance solely to AI.

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

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

How Quickly Returns Will Emerge

It remains unclear when AI investment will produce broad, durable gains in productivity, revenue or margins. Apollo’s aggregate comparisons cannot show whether AI caused the trends, and the cited company examples and experiments do not establish results across the economy. The source material also does not provide a consistent, independently verified measure of returns on AI spending across firms.

The roughly $825 billion figure for Meta is a combination of leases, contractual commitments and possible purchases spread over time, not a single bill due now. It includes items beyond AI, and the timing and eventual use of some capacity may change. The full terms and implications of reported borrowing by AI infrastructure firms, including SpaceX, are also not detailed in the material provided. Forecasts for future capital spending are projections, not completed outlays.

Spending Plans Meet Results

The next evidence will come from company filings, quarterly results and productivity data showing whether spending translates into lower costs, higher output or new revenue. Meta’s 2026 capital-spending outlook is $135 billion to $145 billion; actual spending and the returns on the capacity it funds will become clearer over subsequent reporting periods.

Investors and businesses will also be watching how AI providers finance data centers, chips and cloud services, and whether customers beyond a small group of heavy spenders expand their use. Until company-level outcomes and broader productivity measures show sustained changes, Apollo’s conclusion should be read as a report on what is visible in current aggregates, not a final verdict on AI’s economic impact.

Key Questions

What did Apollo mean by saying AI has had zero payoff outside tech?

Slok was describing aggregate trends: revenue per worker and margins rose for the seven large technology companies Apollo tracked, while the other 493 S&P 500 companies were roughly flat on the measures cited. The analysis does not prove that every non-tech company has seen no benefit from AI.

Does Meta owe $825 billion now?

No. The figure combines future leases, contractual commitments and contingent purchases, many extending over several years. It is not a single immediate payment, and the filing includes some commitments that are not specifically for AI.

Are there examples of AI improving business results?

Yes. The source material cites research showing higher issue resolution per hour in customer support and retail experiments with sales results ranging from no increase to a gain. These examples concern particular tasks and settings, not economy-wide returns.

What evidence would show whether AI is paying off?

Relevant signs would include sustained, attributable changes in productivity, revenue, operating costs or profit margins across more companies. Current aggregate indicators cited by Apollo do not yet show a broad effect, while isolated examples cannot establish one on their own.

Source: ThorstenMeyerAI.com