🔍 Read the full analysis: An Economist Questions AI’s Returns Outside The Tech Sector on ThorstenMeyerAI.com
Get business pricing on monitors, keyboards and dev gear
- Business-only prices and quantity discounts
- Tax-exempt purchasing
- Multiple users, one account, clear invoices
TL;DR
Apollo chief economist Torsten Slok says recent productivity and company data show no visible financial payoff from AI across non-tech companies so far. The conclusion describes aggregate trends, not proof that AI has produced zero gains in every business; some task-level studies report measurable improvements, while large infrastructure bills and financing risks are growing.
Apollo chief economist Torsten Slok said in an October 24 research note that available company and productivity data show no visible financial payoff from AI outside the technology sector, even as firms commit vast sums to computing infrastructure. His conclusion concerns broad averages, not a finding that every non-tech business has gained nothing: task-level studies cited in the source report measurable improvements in some customer-support and retail workflows.
Apollo’s comparison focuses on revenue per worker and profit margins. Its charts show the seven largest technology companies—the so-called Magnificent Seven—at about $780,000 in inflation-adjusted revenue per employee in the third quarter of 2026, while the other 493 S&P 500 companies were around $525,000. Apollo says the latter group’s figure has remained in a similar range since 2021. The technology group’s margins were near 30%, compared with roughly 10% for the rest.
Slok also pointed to San Francisco Federal Reserve productivity measures that, according to Apollo’s reading, show no sign yet of AI lifting productivity growth. Separately, data from Ramp, a corporate spending platform, showed AI-related spending concentrated among a small group of firms: the top 10% accounted for 99.5% of spending on model serving and inference among Ramp’s customers. That dataset covers Ramp customers, not the S&P 500 as a whole.
The apparent gap between investment and broad results comes as infrastructure spending accelerates. Meta’s third-quarter results reported $34.91 billion in capital expenditures and negative free cash flow of $2.12 billion, compared with positive free cash flow of $10.6 billion a year earlier. A filing listed roughly $825 billion in future commitments, including leases and contractual purchases, largely connected to data centers, servers and cloud capacity. The total includes items not exclusively tied 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
AI Spending Meets Uneven Returns
The debate matters because companies are making large, long-term commitments to infrastructure before aggregate evidence shows broad productivity gains. If revenue or output does not rise enough to cover compute, energy and financing costs, spending could weigh on margins or lead firms to slow projects. Slok has described the marginal cost of compute as exceeding the marginal benefit to users, but that is his interpretation of the evidence, not a settled measurement for every company.
The data do not rule out useful returns in specific jobs or firms. The source cites customer-support research reporting 14% more issues resolved per hour, with a 34% gain for novice workers, and retail experiments in which sales results ranged from no increase to a 16.3% increase depending on the workflow. Those findings concern particular applications; they do not establish that similar gains have spread through the wider economy.
For business leaders and workers, the distinction is practical: investment does not itself prove productivity. Companies still need to show whether AI changes the cost, speed or quality of work enough to justify the expense. For investors and lenders, the scale of planned infrastructure raises questions about cash flow, borrowing and whether expected demand will support those commitments.
Products Worth Considering
Samsung 27-Inch Odyssey QD-OLED G8 4K 240Hz Gaming Monitor
Apple iPad Air (4th Generation, 10.9-inch, Wi-Fi, 256GB) – Space Gray
Amazon Kindle Paperwhite (12th Generation, 2024), 16 GB, 7-inch Display, Jade
As an affiliate, we earn on qualifying purchases.
What Apollo’s Data Covers
Apollo’s October 24 “Daily Spark” note, titled “No Signs of AI Payoff Outside the Tech Sector,” drew on several measures rather than a direct experiment that isolates AI’s effect. Revenue per worker and profit margins compare company groups; productivity statistics track the wider economy; Ramp’s figures describe spending by its own business customers. These measures can indicate trends, but they do not establish that AI caused every difference—or that AI had no effect where averages stayed flat.
The source also describes concentrated investment. It says 80.4% of the top 1% of AI spenders in Ramp’s data were venture-backed, compared with 13.1% of the median spender. Slok has called this venture-funded token spending “circularity,” arguing that spending by companies funded to build AI products may not show broad customer demand. The figures are specific to the platform’s customer base and should not be treated as a census of corporate AI use.
Meta’s commitments illustrate the size of the buildout but are not equivalent to a single cash bill due now. Its filing includes 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 source notes that not all of these obligations are AI-related. Separately, forecasts for spending by major cloud and technology companies are projections, not booked commitments.
“has delivered zero financial payoff to corporate America outside of tech”
— Torsten Slok, Apollo chief economist, in an October 24 research note
What the Averages Cannot Show
Apollo’s charts do not isolate AI from other forces affecting company performance, including labor costs, interest rates and tariffs. A flat average across 493 companies could also conceal businesses that gained and others that lost. The data support the narrower conclusion that broad gains are not yet visible in these measures; they do not prove a zero return for every firm.
It remains unclear how much of Meta’s listed commitments will be spent, when the related facilities and capacity will be used, or what share is specifically for AI. The source says Meta’s free cash flow fell sharply in the quarter, but one quarter does not establish a lasting trend. It also cites rising insurance costs for some technology-company debt and a $40 billion borrowing by SpaceX to buy AI chips; the long-term ability of these investments to generate returns is not established by the figures provided.
The source reports that 83% of nearly 9,000 executives said AI had produced no productivity impact so far, with an average impact of 0.2% a year. It does not provide enough detail here about the survey’s full methodology or the precise period used for the average. Those results should be read as reported survey findings, not as a definitive measure of AI’s economy-wide effect.
Evidence From Company Results
The next test will be whether companies can show sustained improvements in output, revenue or margins as AI moves from investment plans into routine operations. Investors will also watch whether the infrastructure buildout produces enough paying demand to support its financing costs. The supplied material does not identify a scheduled follow-up from Apollo or a specific date for an economy-wide productivity milestone.
For now, company filings and future productivity data will offer separate checks: filings can show how quickly spending and obligations grow, while productivity measures may reveal whether output per hour changes over time. Neither spending totals nor isolated success stories settle the wider returns question. The scale and durability of gains outside technology remain open.
Key Questions
What did Apollo’s economist conclude about AI returns?
Torsten Slok said Apollo’s data show no visible financial payoff from AI outside tech in broad company and productivity measures so far. That is an interpretation of aggregate evidence, not proof that every non-tech company has received zero benefit.
Does the report say AI has no value in non-tech businesses?
No. The source cites task-level examples, including higher customer-support resolution rates and retail experiments with sales results ranging from no gain to a 16.3% increase. Those results are specific to the workflows studied and do not establish economy-wide returns.
What does Meta’s roughly $825 billion figure represent?
It combines future lease obligations, contractual commitments and contingent cloud purchases reported in Meta’s filing. The figure is not all due immediately, and the source says not all of it is specifically for AI.
Why are analysts focused on AI financing?
Large infrastructure plans require substantial capital before their long-term returns are clear. The source points to a $40 billion SpaceX borrowing for AI chips and rising insurance costs on some technology debt, but the eventual returns and financing conditions remain uncertain.
Source: ThorstenMeyerAI.com
