AI Investment Across Industries: What The Numbers Really Show

🔍 Read the full analysis: AI Investment Across Industries: What The Numbers Really Show on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get the latest gadgets delivered free with Prime

• Fast, free delivery on millions of items
• Prime Video, Amazon Music and more included
• Member-only deals all year

Start your free Prime trial →Free trial for eligible customers · Cancel anytime

As an affiliate, we earn on qualifying purchases.Hide

Apollo chief economist Torsten Slok’s October 24 analysis found no clear aggregate financial payoff from artificial intelligence outside major technology companies, even as companies make large commitments to data centers, servers and cloud capacity. The finding concerns broad economic and corporate measures, not every AI project: reported workplace studies show gains in some tasks, while the overall return on investment remains difficult to isolate.

Apollo’s charts compare the seven largest technology companies—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—with the other 493 companies in the S&P 500. In Apollo’s reading, inflation-adjusted revenue per worker and profit margins rose for the seven, while the measures for the other 493 remained broadly flat. Its note also cites San Francisco Federal Reserve productivity measures that, it says, show no visible AI-driven boost to productivity growth.

The spending figures point to a widening gap between investment and demonstrated returns. Meta reported $34.91 billion in capital expenditures in the third quarter and narrowed its 2026 capital spending outlook to $135 billion to $145 billion. Its operating cash flow was $32.79 billion, leaving free cash flow at negative $2.12 billion for the quarter, compared with positive $10.6 billion a year earlier. A filing lists roughly $825 billion in leases and contractual commitments, largely connected to infrastructure and cloud capacity; the total includes obligations that extend over time and is not all AI spending.

There are reported gains in particular uses. Source material cites a customer-support study in which issues resolved per hour increased 14% overall and 34% for novice workers, and retail experiments with sales changes ranging from zero to 16.3%, depending on the workflow. It also reports that 83% of nearly 9,000 executives saw no productivity impact so far, with an average estimated effect of 0.2% a year. These figures describe different studies and measures; they do not establish a single economy-wide return.

The Gap Between AI Costs and Returns

The findings matter because AI infrastructure requires major, often long-term spending, while companies’ productivity and financial accounts have yet to show broad gains outside the technology leaders. If returns remain concentrated in a small number of firms or processes, other businesses may face costs without an immediate measurable improvement in output or margins.

That does not mean the investment has no value. New infrastructure may support future services, and gains in a specific workflow may be obscured in company-wide averages. But investors, executives and customers need to distinguish spending, adoption and promised potential from realized financial results. A growing capital bill by itself is not evidence that AI is producing returns.

What Apollo’s Comparison Measures

Apollo’s October 24 note, titled “No Signs of AI Payoff Outside the Tech Sector,” brought together comparisons of revenue per worker and margins, productivity indicators, and spending data from Ramp, a corporate card and spending platform. According to the source material, Ramp’s dataset covers more than 85,000 U.S. business customers. It found that a small share of firms accounted for most tracked spending on model inference, neocloud services and products from OpenAI and Anthropic. Those platform data describe Ramp customers, not the S&P 500 as a whole.

The headline claim followed news coverage of the Apollo note on October 25. The same source material reports that Meta’s third-quarter results were released October 21 and its filing was made October 22. The filing’s roughly $825 billion total combines 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 obligations have different terms and uses, so the headline total should not be read as a bill due immediately or as a pure AI expenditure.

“Artificial intelligence has delivered zero financial payoff to corporate America outside of tech.”

— Torsten Slok, Apollo chief economist, in the October 24 Daily Spark note

What the Aggregate Data Cannot Show

A flat average for 493 companies cannot establish that each company has seen no benefit: gains at some firms could be offset by losses or costs elsewhere. The comparisons also cannot isolate AI’s effect from changes in interest rates, labor expenses, tariffs or other business conditions. Apollo’s evidence supports the narrower conclusion that no broad payoff is visible in the measures it examined; it does not prove that AI caused no improvement in any company.

Other limits remain. Ramp’s spending patterns are not a representative measure of all U.S. firms, and executive surveys rely on reported assessments. The source material does not provide enough detail to compare the customer-support and retail studies by sample size, duration or methodology. It is also unclear how much of Meta’s commitments will be used specifically for AI, how quickly the obligations will become payments, and what returns the infrastructure will generate.

Evidence to Watch in Upcoming Results

Future company filings and earnings reports should show whether heavy spending turns into higher output, revenue or margins, and how much cash companies need to finance infrastructure commitments. Investors will also watch whether credit costs change as AI-related borrowing expands. The figures cited here include reported debt-insurance costs and a large borrowing by SpaceX, but do not establish that financing pressure has reached every AI developer or infrastructure provider.

For now, the clearest test is whether companies can connect particular AI deployments to measurable improvements after accounting for their costs. Until more results accumulate, the evidence supports two points at once: AI investment is substantial, while economy-wide financial returns remain unproven.

Key Questions

Does Apollo’s analysis prove that AI has produced no returns?

No. It finds no clear aggregate payoff in the measures it examines outside major technology companies. Company averages can conceal gains and losses, and the analysis does not isolate AI from other economic factors.

What does Meta’s roughly $825 billion commitment figure represent?

It combines leases not yet started, contractual commitments and contingent cloud purchases reported in Meta’s filing. The total spans future obligations, is not all AI spending and is not an amount due immediately.

Are there measured examples of AI improving work?

The source material cites studies reporting higher customer-support issue resolution per hour and sales changes in some retail experiments. Results vary by workflow, and the supplied information does not give enough methodological detail to treat them as a general forecast.

What evidence would show that AI investment is paying off?

Companies would need to report improvements such as higher output, revenue or margins that can be linked to AI use and compared with its costs. Investment totals alone do not demonstrate a financial return.

Source: ThorstenMeyerAI.com