🔍 Read the full analysis: AI’s Expanding Spend Meets A Narrow Payoff on ThorstenMeyerAI.com
Get business pricing on office and shipping supplies
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Apollo economist Torsten Slok says broad corporate data show no clear financial payoff from AI outside large technology companies, while investment in computing capacity continues to rise. The findings describe aggregate trends, not proof that AI has no benefits: measured gains have appeared in some specific workplace tasks, and the causes of company-level results remain difficult to isolate.
Apollo economist Torsten Slok says available company and productivity data show no clear financial payoff from AI across most of corporate America, even as technology companies and other investors commit heavily to data centers, servers and computing capacity. His October 24 analysis contrasts rising revenue per worker and profit margins at seven major technology companies with largely flat measures for the other 493 S&P 500 companies; the comparison shows an aggregate gap, not proof that AI caused it.
Apollo’s note, titled “No Signs of AI Payoff Outside the Tech Sector,” draws on five charts. It reports that inflation-adjusted revenue per worker has risen for the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—while the measure for the remaining S&P 500 companies has stayed near its range since 2021. The note also shows margins near 30% for the seven and around 10% for the other companies. Those figures describe differences between company groups; they do not by themselves establish what caused them.
The analysis also points to San Francisco Federal Reserve measures of productivity, which it says show no sign yet of an AI-driven lift in productivity growth. Ramp spending data cited by Apollo indicate that a small share of firms accounts for most observed spending on AI model serving, inference and neocloud services. Ramp has more than 85,000 U.S. business customers, but the figures describe those customers and should not be treated as a census of U.S. companies. Apollo says venture-backed firms make up 80.4% of the top 1% of AI spenders, compared with 13.1% of median spenders.
The investment bill is visible in company filings. Meta reported $34.91 billion in capital expenditures in its third quarter and narrowed its 2026 capital-spending outlook to $135 billion to $145 billion. Its operating cash flow was $32.79 billion, while free cash flow was negative $2.12 billion, compared with positive $10.6 billion a year earlier. The company’s 10-Q lists about $355.73 billion in leases not yet started and $454.52 billion in non-cancelable contractual commitments, as well as up to $14.72 billion in contingent cloud-capacity purchases. Together, those categories total roughly $825 billion, though they include commitments extending over years and are not all exclusively for AI.
Financing is another pressure point. The source report says SpaceX borrowed $40 billion to buy AI chips and reports that insurance costs on debt from Meta and CoreWeave have risen. It also cites forecasts for capital spending by five large technology companies—Amazon, Alphabet, Meta, Microsoft and Oracle—of nearly $780 billion in 2026 and more than $1 trillion annually from 2027 through 2030. Those are projections, not booked commitments or completed spending.
Infographic 01 · The “zero” in numbers
What “zero financial payoff” actually measures.
Apollo’s headline rests on five charts. Here are their numbers, and what they can and cannot prove.
Who pays for AI? Share of spending by the top 10% of firms
Ramp customer data via Apollo · non-AI categories in grey for comparison
Spending is concentrated in every category, not only AI; AI is the most extreme.
What the five charts do and do not show
shown not shown
- Shown: revenue per worker and margins rising for the 7 tech giants, flat for the other 493
- Shown: no AI lift yet in US productivity statistics (San Francisco Fed TFP, output per hour)
- Shown: AI spending sits in few, mostly venture-funded firms (Ramp, 85,000+ US businesses)
- Not shown: that AI caused the gap; rates, tariffs and wages move the same lines
- Not shown: gains inside single functions, which can cancel out in an average
- Not shown: the S&P 493 itself in the Ramp data; those are Ramp’s customers
“The AI story will be decided in the S&P 493, not the Mag 7.”Torsten Slok, Apollo chief economist, 24 Oct 2026 (opinion)
Sources: Apollo Daily Spark (24 Oct 2026) with charts by Bloomberg, Macrobond, SF Fed, BLS and Ramp; Apollo (20 Oct 2026); Ramp AI Index. Values with ≈ are read off Apollo’s charts. Assessment: Thorsten Meyer.
Powered by Thorsten Meyer AI
Spending Outruns Visible Returns
The contrast matters because AI infrastructure requires large, often long-term outlays, while the business gains remain uneven and difficult to measure. If companies cannot translate computing costs into higher output, revenue or lower expenses, investment may weigh on margins or raise questions about how projects are funded. Meta’s negative free cash flow in the reported quarter and its extensive future commitments illustrate the scale of that exposure, but one quarter does not establish a lasting financial trend.
For workers and customers, the aggregate figures do not settle whether AI is useful in particular jobs. The source material cites studies in which customer-support workers resolved 14% more issues per hour on average, with a 34% gain among novice workers, and retail experiments ranging from no sales increase to a 16.3% increase depending on the workflow. Those results point to possible gains in defined settings, not a general return that can be assumed across companies.
The central distinction is between spending on AI and earning a return from using it. Apollo’s evidence suggests the benefits are not yet visible in broad corporate averages. It does not show that every company has received no benefit, nor does it isolate AI from other factors affecting revenues, costs and productivity.
Products Worth Considering
FLEXISPOT 42-Inch Electric Standing Desk Converter
FLEXISPOT 32-Inch Electric Standing Desk Converter with Keyboard Tray
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
As an affiliate, we earn on qualifying purchases.
What Apollo’s “Zero” Measures
Slok’s phrase that AI has delivered “zero financial payoff” outside technology is an interpretation of aggregate trends, rather than a direct measurement of AI’s effect on each company. As presented in the source material, the comparison is between seven large technology businesses and the other 493 companies in the S&P 500. A group average can conceal companies that are gaining or losing, and it cannot separate AI’s effects from changes in interest rates, tariffs, labor costs or other business conditions.
The timing also matters. Apollo published its note on October 24, after reporting on the analysis by outlets including TechCrunch, PYMNTS and Bloomberg. Meta’s third-quarter release came on October 21 and its 10-Q on October 22. The source report says Meta shares initially rose after results and then closed 6.1% lower the following day; it also says the company’s five-year credit default swaps reached a record 114.79 basis points, citing Fortune. Those market movements are evidence of investor reaction, not proof that the commitments will cause financial distress.
The source material further cites a survey of nearly 9,000 executives in which 83% reported no productivity impact from AI so far and the average estimated impact was 0.2% a year. Such survey results reflect reported experience and depend on how productivity and AI use are defined. They add to the picture of limited broad gains to date, but do not rule out later effects.
The Returns Remain Hard to Isolate
It is not yet clear how much AI caused the reported differences in company revenue per worker, margins or productivity. The cited group comparisons do not control for other economic factors, and aggregate figures may hide strong results in some firms alongside weak results in others. Ramp’s spending data also cover its own business customers, not the entire corporate sector.
The $825 billion figure for Meta combines leases, contractual commitments and contingent purchases over multiple years; it is not a bill due at once, and the source says some commitments are not AI-related. The source material does not provide a full breakdown of when each amount will be paid or what returns each project is expected to generate. It also does not establish the terms or repayment schedule of SpaceX’s reported $40 billion borrowing.
Finally, the cited task-level gains come from particular studies or experiments. The source does not specify enough about their methods, duration or costs to treat them as evidence of economy-wide returns. Whether those gains persist and scale across different workflows remains unresolved.
Watch Spending and Productivity Data
The next evidence will come from company results, capital-spending plans and updated productivity measures. Investors and businesses will be watching whether the large technology companies can turn infrastructure investment into revenue growth, improved margins or measurable savings, and whether those gains spread beyond firms that sell AI services.
For Meta, subsequent filings and cash-flow reports can clarify how commitments develop into actual spending and whether free cash flow recovers. For the wider market, disclosures from cloud providers and AI companies may offer more detail on financing needs and customer demand. Apollo’s October analysis is a snapshot of reported conditions, not a forecast that returns will remain absent.
Key Questions
What did Apollo conclude about AI’s financial payoff?
Apollo economist Torsten Slok said broad data show no clear financial payoff outside major technology companies. The statement reflects aggregate comparisons and does not prove that no individual company or task has benefited.
Does the analysis prove AI has produced no productivity gains?
No. Apollo cited productivity measures showing no clear AI-driven lift so far, but those measures do not isolate AI’s effects at every company or job. The source also describes task-level studies with reported gains in some customer-support and retail workflows.
What does Meta’s roughly $825 billion figure include?
It combines about $355.73 billion in leases not yet started, $454.52 billion in non-cancelable contractual commitments and up to $14.72 billion in contingent cloud-capacity purchases. These obligations span multiple years, and the source says they are not all exclusively for AI.
Why is AI spending raising financing concerns?
Building data centers and buying servers and chips require large amounts of capital before returns are certain. The source reports a $40 billion SpaceX borrowing for AI chips and rising insurance costs on debt from Meta and CoreWeave, while the precise financing terms and long-term returns remain unclear.
Source: ThorstenMeyerAI.com
