🔍 Read the full analysis: The Real Numbers On AI Investment And Returns Across Industries on ThorstenMeyerAI.com
TL;DR
Apollo economist Torsten Slok says aggregate data show no clear financial payoff from AI outside the technology sector, while corporate commitments to computing infrastructure keep growing. The evidence does not prove AI has produced no gains: some task-level studies report benefits, and broad averages can mask differences among companies and jobs.
An October 24 analysis by Apollo chief economist Torsten Slok found no visible improvement in broad productivity and financial measures for most large US companies outside technology, sharpening questions about the returns on a growing wave of AI infrastructure spending. The analysis describes aggregate trends, not proof that AI has generated no benefits: selected workplace studies show gains in specific tasks, while the scale and timing of returns across industries remain uncertain.
Apollo’s note, titled “No Signs of AI Payoff Outside the Tech Sector,” compares the seven largest technology companies, commonly called the Magnificent Seven, with the other 493 companies in the S&P 500. Its charts show revenue per worker and profit margins rising among the seven, while the corresponding figures for the other 493 remain broadly flat. Apollo also cites San Francisco Federal Reserve productivity measures that it says show no sign yet of AI boosting productivity growth.
The figures are not a direct measurement of AI’s effect. Differences between technology firms and other companies can reflect business models, costs, market conditions and other factors, as well as AI use. A flat average can also conceal companies or teams that gained while others did not. Apollo’s conclusion is best read as a finding that aggregate returns are not yet apparent, rather than evidence that every non-technology company has received no benefit.
At the same time, infrastructure bills are growing. Meta reported $34.91 billion in capital expenditure in its third quarter and narrowed its 2026 spending forecast to $135 billion to $145 billion. Its October 22 filing lists roughly $355.73 billion in leases that have not started and $454.52 billion in non-cancelable contractual commitments, plus up to $14.72 billion in contingent cloud purchases. These amounts include commitments extending over several years, and not all relate to AI.
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 Spending and Returns
The mismatch matters because companies, investors and customers are trying to determine whether AI investment will translate into durable productivity or revenue gains. Large infrastructure commitments can support future services and computing capacity, but they also create costs before the eventual returns are known. Meta’s reported negative $2.12 billion in third-quarter free cash flow, compared with positive $10.6 billion a year earlier, illustrates the near-term financial pressure alongside its investment plans. That figure is for the company as a whole and does not isolate AI spending as the cause.
Evidence from specific uses is more encouraging but narrower. The source material cites research finding customer-support workers resolved 14% more issues per hour, with a 34% increase for novice workers, and retail experiments reporting sales changes ranging from zero to 16.3%, depending on the workflow. Those results suggest that gains may depend on how a tool is applied, not simply whether a company buys access to a model. They do not establish that the same effects will recur across firms or industries.
The distinction affects decisions about hiring, software budgets and infrastructure. If AI improves a defined task, a company may be able to measure the time saved or output produced. If spending is concentrated among a small group of firms without clear productivity gains elsewhere, businesses may face higher costs before their use cases mature. The current evidence supports caution about broad promises, not a conclusion that investment cannot pay off.
Popular right now
Breville Bambino Espresso Machine, Brushed Stainless Steel
Breville Barista Pro Espresso Machine with Grinder, Brushed Stainless Steel
Ninja Luxe Café Mini Plus Espresso and Drip Coffee Machine with AutoFroth, Stainless Steel, ES351
As an affiliate, we earn on qualifying purchases.
What Apollo’s Charts Compare
Apollo’s analysis draws on five sets of measures: inflation-adjusted revenue per worker, profit margins, productivity indicators, company spending on AI services and the financing of heavy spenders. It reports that the Magnificent Seven were near $780,000 in revenue per employee in the third quarter of 2026, compared with about $525,000 for the other S&P 500 companies. The seven’s margins were around 30%, while the rest were around 10%. These are group-level figures, not estimates of the amount AI contributed.
For spending patterns, Apollo cites data from Ramp, a corporate card and spending platform with more than 85,000 US business customers. Its data show 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. Those shares describe Ramp’s customer data and should not be treated as a census of all US businesses.
The spending question has also moved beyond chip purchases to financing and long-term capacity. Meta’s commitments cover items such as data centers, energy infrastructure, network capacity and cloud services; the company’s filing includes obligations that run through 2036. The reported totals are commitments over time, not an amount paid in a single year. Separate forecasts for major technology companies’ future capital spending are projections, not booked obligations.
„No Signs of AI Payoff Outside the Tech Sector“
— Torsten Slok, Apollo chief economist, in the October 24 note
What the Aggregate Data Cannot Show
The available figures do not isolate AI as the cause of changes in company revenue, margins or productivity. They cannot show whether an individual firm’s results reflect AI, interest rates, labor costs, tariffs or other factors. Nor does the difference between technology firms and other S&P 500 companies establish that AI caused the gap.
It is also unclear how representative the task-level studies are across workplaces. Reported customer-support and retail results apply to particular experiments and workflows; they do not establish economy-wide effects. The source material says 83% of nearly 9,000 executives reported no productivity impact so far and gives an average of 0.2% a year, but does not supply enough study detail here to assess the survey’s methods, timing or definition of productivity impact.
Meta’s commitment total is not all AI-related, and it is not the same as cash already spent. The cited material also does not establish how quickly companies will use the contracted capacity, what revenue it may generate, or whether future productivity measurements will capture gains that are now difficult to quantify.
Evidence to Watch in Coming Reports
The next test will be whether company filings and productivity data show sustained gains alongside continued investment. Investors and business leaders will be watching for clearer links between AI use and revenue, margins, output per worker or measurable task improvements, rather than relying on spending levels or executive expectations alone.
Upcoming results from major technology companies may clarify how much they are spending and how infrastructure plans are developing. Broader productivity measures and more detailed studies of workplace deployments will be needed to assess whether task-level benefits spread across industries. For now, Apollo’s analysis documents a gap between visible investment and aggregate returns, while leaving open whether those returns will emerge later or remain concentrated in particular uses.
Key Questions
Did Apollo prove that AI produces no financial returns?
No. Apollo reported that broad measures show no visible payoff outside technology so far. Its charts do not isolate AI’s effects or rule out gains at individual companies and in specific tasks.
What does Meta’s roughly $825 billion commitment figure mean?
It combines reported leases that have not started, non-cancelable contracts and contingent cloud purchases. The commitments extend over time, include items not all tied to AI, and are not $825 billion already spent.
Are there measured workplace gains from AI?
The source material cites task-specific findings, including a 14% increase in customer-support issues resolved per hour and retail experiment results ranging from no sales increase to 16.3%. These findings do not establish results across all companies.
Why can’t the S&P 500 comparison settle the question?
The comparison is between groups of companies, not a controlled test of AI use. It cannot separate AI from other causes of revenue, margin or productivity changes, and group averages may hide gains and losses among individual firms.
Source: ThorstenMeyerAI.com
