🔍 Read the full analysis: AI’s Growing Price Tag: What An Economist Says About The Payoff on ThorstenMeyerAI.com
Get hardware and tech essentials delivered free with Prime
- Fast, free delivery on millions of items
- Prime Video, Amazon Music and more included
- Member-only deals all year
TL;DR
Apollo chief economist Torsten Slok says aggregate data show no clear AI-driven financial payoff outside the technology sector, even as companies commit heavily to infrastructure. The figures describe broad trends, not proof that AI creates no value: targeted trials show gains, while the timing and scale of returns remain uncertain.
Apollo chief economist Torsten Slok said October 24 that available aggregate data show no visible financial payoff from AI outside the technology sector, even as companies make large commitments to data centers, servers and cloud capacity. His conclusion highlights a widening gap between measurable investment and returns that are not yet apparent across most businesses; it does not establish that AI has produced no gains in individual companies or work processes.
Slok’s Apollo “Daily Spark” note, titled “No Signs of AI Payoff Outside the Tech Sector,” compared revenue per worker and profit margins for the seven largest technology companies with those of the other 493 S&P 500 firms. Apollo’s charts show revenue per worker and margins rising for the seven, while the wider group’s figures have remained broadly flat. The note also cited San Francisco Federal Reserve productivity measures, which it said show no signs of AI boosting productivity growth.
That is evidence about aggregate trends, not a direct measurement of AI’s effect on each company. A flat average can conceal firms that gained and firms that did not, and the data cannot isolate AI from other forces such as labor costs, tariffs or interest rates. Slok’s “zero” wording is an interpretation of the pattern, rather than a finding that every non-tech company has received no benefit.
The spending is more directly visible in company disclosures. Meta’s October 22 filing listed about $355.73 billion in leases not yet started and $454.52 billion in non-cancelable contractual commitments, mostly tied to cloud capacity, servers and network equipment, plus up to $14.72 billion in contingent cloud purchases. Together, those items total roughly $825 billion. They extend into future years, include commitments that are not all AI-related, and are not equivalent to cash already spent.
Meta reported $34.91 billion in capital expenditures in the third quarter and narrowed its 2026 spending forecast to $135 billion to $145 billion. Operating cash flow of $32.79 billion was below that quarter’s capital expenditures, leaving free cash flow at negative $2.12 billion, compared with positive $10.6 billion a year earlier. The company’s quarterly revenue was $64.60 billion, up 26% year over year, so the figures show both business growth and a heavier infrastructure bill.
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
The Gap Between AI Spending and Returns
The difference between large investment and uncertain returns matters for companies, investors and users because AI infrastructure requires ongoing spending on computing capacity, energy and related services. If adoption does not lift revenue or productivity enough to cover those costs, the investment can pressure margins and make future expansion harder to fund. The available evidence does not establish that this outcome will occur, but it shows why companies and lenders are scrutinizing the payoff.
There are also signs that gains can appear in specific uses even when broad productivity data do not show them. The source material cites a customer-support study in which issues resolved per hour rose 14% overall and 34% for novice workers. It also reports retail experiments ranging from no sales increase to a 16.3% increase, depending on the workflow. These results are narrower than economy-wide productivity measures and should not be treated as representative of all firms.
A survey of nearly 9,000 executives found that 83% reported no productivity impact so far, while the reported average gain was 0.2% a year. Those figures describe respondents’ assessments and do not settle whether gains may emerge later. For workers and customers, the practical question is whether AI tools improve a defined task enough to justify their direct and indirect costs.
Products Worth Considering
Trezor Safe 7 Crypto Hardware Wallet
Apple MacBook Pro 14.2-inch with M3 Max, 14-Core CPU, 30-Core GPU, 36GB RAM, 1TB SSD, Space Black
Trezor Safe 5 Crypto Hardware Wallet with Color Touchscreen, Black Graphite
As an affiliate, we earn on qualifying purchases.
What Apollo’s Measures Capture
Apollo’s comparison focuses on differences between the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—and the remaining S&P 500 companies. It also draws on Ramp spending data from more than 85,000 U.S. business customers. Apollo reported that the top 10% of firms accounted for 99.5% of model-serving and inference spending, 99.0% of neocloud spending and 95.6% of spending on OpenAI and Anthropic. The figures indicate concentrated spending among Ramp customers; they do not describe every U.S. company.
The commitment figures and cash-flow results answer different questions. Meta’s quarterly cash-flow statement tracks money received and paid during a period. Its filing’s leases and contracts describe future obligations across multiple years. Combining those obligations gives a sense of the scale of planned capacity, but not a single-year expense or a confirmed measure of AI investment alone. Meta’s commitments include non-AI items, including Reality Labs hardware.
Slok has also described token spending as a cost center and argued that the marginal cost of compute may exceed the benefit delivered to users. Those are his economic assessments, not independently established outcomes for every AI service. The central distinction is between the scale of investment, which filings make visible, and the returns, which remain difficult to identify across broad company data.
“No Signs of AI Payoff Outside the Tech Sector”
— Torsten Slok, Apollo chief economist, in the October 24 Daily Spark note
It remains unclear how much of the flat performance among non-tech companies reflects AI’s limited adoption, how much reflects delayed gains, and how much is driven by unrelated economic pressures. Aggregate revenue, margins and productivity measures cannot attribute changes to AI by themselves. Apollo’s analysis also does not provide a company-by-company account of AI costs and resulting revenue.
The source material reports that SpaceX borrowed $40 billion to buy AI chips, but it does not provide enough detail here to assess the loan’s terms, repayment plan or expected returns. Nor does the material establish how much of Meta’s roughly $825 billion in commitments will ultimately be used for AI, how much may change, or what revenue those commitments will generate. Spending projections for large technology companies are forecasts, not recorded expenditures or guarantees of future returns.
The Next Test Is Measurable Returns
The next evidence will come from company filings and operating results: whether infrastructure spending translates into sustained revenue growth, improved productivity or lower costs, and whether those gains exceed the cost of computing and financing. Meta’s later quarterly reports can show how capital spending and cash flow develop against its 2026 forecast. Similar disclosures from other major technology companies will help clarify whether the buildout is producing returns beyond the firms selling AI infrastructure.
For businesses adopting AI, more useful evidence may come from controlled comparisons within specific tasks, such as support response times or sales outcomes, with costs and time periods reported alongside the gains. Until such results accumulate, the broad conclusion remains limited: Apollo sees no aggregate payoff outside tech so far, while some focused applications show measurable benefits. The duration and reach of those benefits are still unsettled.
Key Questions
What did Apollo’s economist conclude about AI’s payoff?
Torsten Slok’s October 24 note said aggregate data show no visible financial payoff outside the technology sector. That is his interpretation of broad trends, not proof that no individual non-tech company has gained from AI.
Does Meta already owe $825 billion in cash?
No. The roughly $825 billion figure combines future leases, contractual commitments and contingent purchases disclosed in Meta’s filing. It spans multiple years, is not all due immediately, and includes items that are not all AI-related.
Are there measured examples of AI improving work?
Yes. The source material cites a customer-support study reporting 14% more issues resolved per hour overall and 34% for novice workers. Retail experiments showed results ranging from no sales increase to a 16.3% increase, depending on the workflow.
What is not yet known about AI’s economic impact?
It is not yet clear how widely task-level gains will spread, whether they will persist, or whether they will outweigh computing and financing costs. Broad economic measures also cannot separate AI’s effects from other changes in the economy.
Source: ThorstenMeyerAI.com
