🔍 Read the full analysis: AI Spending Keeps Exploding, But Where Are The 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 recent aggregate data show no clear financial payoff from AI outside the technology sector. At the same time, Meta disclosed about $825 billion in future contractual and lease commitments, mostly tied to infrastructure, while research finds measurable gains in some specific tasks. The aggregate data do not prove AI has produced no returns at individual companies.
Apollo chief economist Torsten Slok said recent data show no clear financial payoff from AI outside the technology sector, even as companies commit vast sums to data centers, chips and cloud services. The debate sharpened after Meta’s filings disclosed about $825 billion in future lease and contractual commitments, much of it for infrastructure, though the figure is not all AI spending and does not establish that the investment has failed.
In an October 24 Apollo note, Slok compared revenue per worker and profit margins at the seven largest technology companies with those of the other 493 S&P 500 firms. Apollo’s charts show the technology group’s revenue per employee and margins rising, while the broader group’s measures have been comparatively flat. Slok also cited San Francisco Federal Reserve productivity measures that, in his reading, show no visible AI boost to productivity growth.
Those charts describe broad economic and company trends, not a direct experiment measuring AI’s effect. A flat average across hundreds of businesses can mask gains at some firms and losses at others, and the figures cannot isolate AI from other influences such as labor costs, interest rates or tariffs. Ramp spending data cited by Apollo also cover its business customers, rather than the S&P 500 as a whole; the note says AI spending is concentrated among a small share of firms.
Meta’s October 22 quarterly filing lists about $355.73 billion in leases that have not yet started and about $454.52 billion in non-cancelable contractual commitments, plus up to $14.72 billion in contingent cloud purchases. The commitments include data centers, energy and network infrastructure, servers and third-party cloud capacity, and are scheduled across multiple years. Meta’s third-quarter release separately reported $34.91 billion in capital expenditures and negative free cash flow of $2.12 billion. The company forecast capital spending of $135 billion to $145 billion for 2026.
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 Cost of Building AI Capacity
The gap between investment and demonstrated returns matters because AI infrastructure spending is no longer confined to a handful of chip purchases. Large technology companies are planning facilities and long-term capacity contracts that create bills extending years into the future. If revenue or productivity gains arrive slowly, firms may face pressure on cash flow, margins and their ability to finance more construction.
That risk should not be mistaken for proof that the spending is wasteful. Meta’s commitments are not equivalent to cash already paid, and some cover non-AI operations. The company also reported 26% year-over-year revenue growth in the third quarter, according to its release. Investors must weigh current business performance against the scale and duration of future obligations, rather than treating a single aggregate productivity measure as a verdict on AI.
The issue also reaches companies buying AI tools. Slok argued that token and compute costs can weigh on margins if they do not produce enough additional value. For employers and customers, the practical question is whether AI improves a defined process enough to justify its costs, not whether the technology has already lifted economy-wide productivity statistics.
Products Worth Considering
Jackery Explorer 2000 v2 Portable Power Station with Two 200W Solar Panels
Jackery Explorer 300 Portable Power Station, 292Wh
Sonos Era 100 Wireless Smart Speaker, Black
As an affiliate, we earn on qualifying purchases.
Aggregate Trends, Specific Gains
Apollo’s note drew attention to a distinction between companies that sell AI-related products and services and businesses adopting them. Its comparison covers the “Magnificent Seven”—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—and the rest of the S&P 500. The charts show stronger recent trends for the seven, but they do not establish that AI alone caused the difference.
Other research cited in the source material points to measurable gains in particular workflows. A customer-support study found 14% more issues resolved per hour, with a larger gain among novice workers. Retail experiments reported sales changes ranging from no increase to 16.3%, depending on the workflow. These results concern specific settings; they do not show that every company will see similar returns.
In a separate survey of nearly 9,000 executives cited by the source, 83% reported no productivity impact from AI so far, and the reported average gain was 0.2% a year. Those are survey findings, not a comprehensive measure of all firms or a causal estimate. Together, the examples suggest that benefits can be visible in targeted applications even while broad productivity data remain inconclusive.
“No Signs of AI Payoff Outside the Tech Sector”
— Torsten Slok, Apollo chief economist, in an October 24 Daily Spark note
What the Spending Figures Cannot Show
Apollo’s “zero” is an interpretation of group-level trends, not a finding that no company outside tech has gained from AI. The available comparisons do not isolate the effect of AI from other economic factors, and averages can conceal differences among companies, departments and tasks. The source material also gives no full methodology for independently evaluating every cited productivity and business survey figure.
Meta’s roughly $825 billion total combines future lease obligations, contractual commitments and contingent cloud purchases. It is not an amount already spent, is spread over time and includes items that are not all AI-related. The source material does not provide a complete breakdown of how much will ultimately be used for AI, or how the company expects each commitment to generate revenue or savings. It also does not establish that Meta’s negative third-quarter free cash flow was caused solely by AI investment.
The broader financing picture remains unsettled. The source reports a $40 billion SpaceX borrowing for AI chips and concerns about the cost of financing infrastructure, but the supplied material ends before presenting the full details of the financing developments it introduces. Predictions about future industry consolidation or spending levels should be treated as forecasts, not settled outcomes.
Returns Will Be Tested in Results
The next evidence will come from company results and updated filings: whether revenue growth, operating margins, cash flow or output per worker improve as AI spending continues. For Meta, investors will be able to compare its later capital expenditure and cash-flow reports with the 2026 outlook and the commitments disclosed in its quarterly filing.
Businesses adopting AI will also need to report results at the level of specific tasks if they want to show a return: costs before and after implementation, output, quality and the time period measured. Until those results are more widely available, Apollo’s charts can signal a lack of broad visible gains, but cannot settle the return question for individual companies. The central test remains whether measurable benefits grow faster than the cost of the systems and capacity being bought.
Key Questions
Did Apollo prove that AI has produced no financial returns outside tech?
No. Apollo’s note interprets broad company and productivity trends. Those data do not directly measure AI’s effect at every business, and they may conceal gains or losses within individual companies.
What does Meta’s roughly $825 billion commitment figure include?
It combines about $355.73 billion in leases that have not yet started, about $454.52 billion in non-cancelable contractual commitments, and up to $14.72 billion in contingent cloud purchases. The obligations extend across future years and are not all AI-related or already paid.
Are there examples of AI producing measurable gains?
Yes. The source cites a customer-support study finding 14% more issues resolved per hour and retail experiments with sales changes ranging from zero to 16.3%, depending on the workflow. These results are specific to the settings studied and do not guarantee similar outcomes elsewhere.
Why is Meta’s free cash flow relevant?
Meta reported negative free cash flow of $2.12 billion in the third quarter, compared with positive free cash flow of $10.6 billion a year earlier, according to the source material. The figure offers a snapshot of cash remaining after investment, but does not by itself identify AI as the sole cause.
Source: ThorstenMeyerAI.com
