🔍 Read the full analysis: The Uneven Economics Of AI: Big Spending, Limited Returns on ThorstenMeyerAI.com
Get business pricing on tech for your team
- 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 visible financial payoff from AI outside major technology companies, though that conclusion is an interpretation of aggregate trends rather than proof that no firms benefit. At the same time, Meta’s filings show about $825 billion in future commitments, while research cited in the source reports measurable gains in some AI-assisted tasks and little effect across most surveyed firms.
Apollo economist Torsten Slok says financial returns from artificial intelligence are not visible across most of corporate America, even as technology companies and other firms commit vast sums to computing infrastructure. The October 24 note, titled “No Signs of AI Payoff Outside the Tech Sector,” contrasts rising revenue per worker and profit margins among seven major technology companies with flat measures for the other 493 S&P 500 firms. It does not establish that no individual businesses are benefiting: the claim is about broad aggregate trends, while the spending commitments show how quickly the cost of the AI buildout is growing.
Apollo’s note compares the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—with the other 493 S&P 500 companies. In Apollo’s charts, inflation-adjusted revenue per worker rises for the seven but remains broadly flat for the rest; profit margins reach about 30% for the seven and stay near 10% for the others. The note also cites San Francisco Federal Reserve productivity measures, which it says show no signs yet of AI lifting productivity growth. Those measures track the economy overall and do not isolate AI as the cause of any change.
The spending picture is more directly documented in company filings. Meta reported $34.91 billion in capital expenditures in the third quarter and narrowed its 2026 capital spending forecast to $135 billion to $145 billion. Its October 22 filing lists about $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. The combined figure is roughly $825 billion in future commitments, much of it for data centers, cloud capacity, servers and network equipment. The total is not all AI spending; the source notes that some commitments cover Reality Labs hardware.
The source also describes financing pressures beyond Meta. SpaceX borrowed $40 billion to buy AI chips, while insurance costs for debt linked to Meta and CoreWeave reportedly rose. These details point to the scale of funding required, but they do not establish that the borrowing itself reflects financial distress or that every planned commitment will become 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
AI’s Bill Outpaces Visible Returns
The gap matters because companies are making long-term infrastructure commitments before broad productivity gains are apparent in economy-wide data. If the returns remain concentrated among a small set of technology firms, businesses outside that group may face pressure to justify AI costs to investors and customers. Large infrastructure bills may also shape how much capacity is built, who can afford to use it and whether providers can earn enough to cover their financing.
But the aggregate picture should not be mistaken for a verdict on every AI project. The source cites studies reporting 14% more customer-support issues resolved per hour, rising to 34% for novice workers, and retail experiments with sales changes ranging from zero to 16.3%, depending on the workflow. It also cites a survey in which 83% of nearly 9,000 executives reported no productivity impact so far, with an average estimated effect of 0.2% a year. These findings use different methods and populations; they illustrate uneven results rather than a single, settled measure of AI’s economic return.
Products Worth Considering
DJI Osmo Mobile 8 Gimbal with Mic Mini Transmitter, Infinity Black
Logitech MX Keys S Wireless Keyboard, Graphite
Samsung 27 Odyssey OLED G5 (G50SF) QHD QD-OLED Gaming Monitor
As an affiliate, we earn on qualifying purchases.
What Apollo’s Comparison Measures
Apollo’s October 24 note was picked up by news outlets the following day. Its headline uses the phrase “zero financial payoff,” but the evidence described in the source is a comparison of broad indicators, not a direct accounting of the profits attributable to AI. Revenue per worker and profit margins can reflect many factors, including business models, labor costs, prices and market conditions. A flat average among 493 companies can also conceal firms whose gains and losses offset one another.
The note points to concentration in AI-related spending as another part of the picture. Data from Ramp, a corporate card and spending platform with more than 85,000 U.S. business customers, show that the top 10% of firms account 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 describe Ramp customers, not the S&P 493 as a whole. Apollo also reports that 80.4% of the top 1% of AI spenders are venture-backed, compared with 13.1% of median spenders.
Slok has separately described token spending as a potential cost center and argued that compute costs can exceed the benefit delivered to users at the margin. Those are his economic assessments, not independently proven conclusions. The source also cites forecasts that five large technology companies could spend nearly $780 billion in 2026 and more than $1 trillion annually from 2027 through 2030; these are projections, not confirmed spending or contractual obligations.
“No Signs of AI Payoff Outside the Tech Sector”
— Torsten Slok, Apollo chief economist, in the October 24 Daily Spark note
How Much AI Is Paying Back
The central uncertainty is whether the weak aggregate productivity picture reflects limited returns, delayed returns or gains that are concentrated in particular firms and tasks. The Apollo measures do not isolate AI’s effect from other economic forces, and the Ramp spending data cover that platform’s customers rather than the entire U.S. corporate sector. The reported task-level studies show potential gains in specific settings but cannot establish how often those results carry over to other workplaces.
It is also unclear how much of Meta’s roughly $825 billion in commitments is directly tied to AI, how much will ultimately be used, and what returns the company expects from each investment. The filing includes obligations for infrastructure and hardware beyond AI. Likewise, the cited spending forecasts may change, while borrowing and credit-insurance reports do not by themselves show whether firms can service future obligations. No single measure in the source settles whether the overall AI buildout will pay for itself.
Evidence to Watch in Upcoming Results
The next tests will come from company results and updated economic data: whether firms outside the largest technology group show sustained productivity, revenue or margin gains, and whether those gains can be linked to AI use rather than other changes. Investors will also watch for revisions to capital spending plans and disclosures about infrastructure commitments, financing and utilization. Meta’s spending outlook and future quarterly cash-flow figures will help show how its investment pace affects the company’s finances.
For now, the evidence supports a narrower conclusion than the headline claim: AI spending is large and concentrated, while broad-based financial returns remain difficult to see. The balance could shift as companies report more results and AI tools move from trials into regular operations, but the timing and scale of any wider payoff remain uncertain.
Key Questions
What did Apollo mean by “zero financial payoff”?
It was Slok’s interpretation of broad corporate indicators, including revenue per worker, profit margins and productivity data. It was not a finding that every company outside the technology sector has earned no return from AI.
What does Meta’s roughly $825 billion figure include?
Meta’s filing lists 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 purchases. The commitments cover infrastructure and hardware, and the source says they are not all AI-related.
Are companies seeing any measurable gains from AI?
The source cites task-level research reporting improved customer-support output and retail results ranging from no sales change to a 16.3% increase, depending on the workflow. It also cites an executive survey in which most respondents reported no productivity impact so far. Results vary by setting and study.
Does the Apollo note prove AI is not productive?
No. The note describes aggregate trends that do not show a broad productivity boost yet. Those measures cannot identify every firm-level result or separate AI’s effects from other economic factors.
Source: ThorstenMeyerAI.com
