Why AI Investment Has Yet To Pay Off In Most Sectors

🔍 Read the full analysis: Why AI Investment Has Yet To Pay Off In Most Sectors on ThorstenMeyerAI.com

Before you orderOffer from Amazon

Get hardware and tech essentials 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 says broad corporate data show no visible financial payoff from artificial intelligence outside major technology companies so far, even as Meta and other firms commit hundreds of billions of dollars to computing infrastructure. The finding describes aggregate trends, not proof that AI produces no gains: productivity measures have yet to show a broad lift, while some studies report improvements in specific tasks.

Apollo’s October 24 note, titled No Signs of AI Payoff Outside the Tech Sector, compared revenue per worker and profit margins for the Magnificent Seven—Alphabet, Amazon, Apple, Meta, Microsoft, Nvidia and Tesla—with the other 493 S&P 500 companies. Apollo’s charts show revenue per employee rising for the seven, to about $780,000 in the third quarter of 2026, while the other companies remained near $525,000, within a range they had occupied since 2021. The seven’s margins rose to roughly 30%; the remaining companies stayed around 10%.

The note also cited San Francisco Federal Reserve productivity measures that, according to Apollo, show no sign yet of an AI-driven boost to productivity growth. Separately, an executive survey cited in the source material found that 83% of nearly 9,000 executives reported no productivity impact from AI so far, with an average reported effect of 0.2% a year. Those measures do not establish that AI caused the difference between the technology giants and other firms, or explain changes in any one company’s performance.

Spending is concentrated among a small group of businesses. Data from Ramp, a corporate spending platform with more than 85,000 U.S. business customers, show the top 10% of firms accounted for 99.5% of spending on model-serving and inference services, 99.0% of neocloud spending, and 95.6% of spending on OpenAI and Anthropic. The figures describe Ramp customers, not all U.S. companies. The source material also reports concrete, task-level gains: a customer-support study found 14% more issues resolved per hour, rising to 34% for novice workers, while retail experiments recorded sales changes from no increase to a 16.3% gain, depending on the workflow.

The Gap Between AI Costs and Returns

The mismatch matters because companies are spending on infrastructure before broad financial returns are evident in the aggregate data. Meta reported $34.91 billion in capital expenditures in the third quarter and narrowed its 2026 spending outlook to $135 billion to $145 billion. Its filing listed about $825 billion in future commitments, including leases and contracts for data centers, energy, cloud capacity, servers and network equipment. The figure is not all AI spending; the source material says it also includes some Reality Labs hardware.

Meta’s quarterly revenue rose 26% to $64.60 billion, but operating cash flow of $32.79 billion was below its capital expenditures, leaving negative free cash flow of $2.12 billion. That compares with positive free cash flow of $10.6 billion a year earlier. One quarter does not settle whether the investment will pay off over time, but it illustrates why investors and companies are weighing infrastructure costs against returns that remain hard to measure across most sectors.

Financing is another pressure point. The source material says SpaceX borrowed $40 billion to buy AI chips and reports that insurance costs for debt issued by Meta and CoreWeave increased. Such developments do not prove that AI investment is unsustainable. They do show that the debate extends beyond whether companies can obtain chips and electricity to whether expected returns can support the cost of building and financing computing capacity.

What Apollo’s Figures Measure

Apollo’s conclusion relies on comparisons across company groups and economy-wide productivity measures. These are useful for tracking whether gains are appearing broadly, but they are not a direct experiment measuring the effect of AI. The 493-company average could conceal firms that benefit and others that lose ground, with the changes offsetting each other. It also cannot isolate AI from other influences, such as labor costs, interest rates or tariffs.

The spending data have limits too. Ramp’s figures cover its own business customers, and the high concentration of purchases does not show whether the biggest buyers are earning a return. Meta’s commitments cover obligations extending over years, not cash already spent in a single period. Forecasts cited in the source material for major cloud and technology companies’ future capital spending are projections, not booked commitments.

Against that broad picture, individual applications can produce measurable results. Customer-support and retail examples suggest that outcomes vary by task and implementation. They do not establish that gains will spread across entire industries or translate into higher company-wide profits. Apollo economist Slok’s argument is that mainstream firms may be treating AI token use as a cost that weighs on margins rather than as a source of growth.

“No Signs of AI Payoff Outside the Tech Sector”

— Torsten Slok, Apollo chief economist, in Apollo’s October 24 note

How Broad AI Returns Remain

The central uncertainty is whether the lack of a broad productivity or margin signal reflects weak returns, a delay before benefits appear, or gains that are confined to particular companies and tasks. Apollo’s comparisons do not isolate AI’s effect, and the reported workflow results cannot establish that similar gains will occur elsewhere.

It is also unclear how much of Meta’s roughly $825 billion in commitments will ultimately be used, how much relates directly to AI, and whether future revenue will justify the spending. The source material reports rising financing costs and a large SpaceX borrowing, but does not provide enough detail to establish that the broader industry is facing a funding shortage or that any named company’s commitments will lead to losses.

Evidence to Watch in Company Results

The next test will be whether companies show measurable returns as they report results and continue spending. Investors will be watching capital expenditure, cash flow, margins and productivity, alongside evidence that AI tools improve specific business processes. Meta’s stated 2026 capital-spending range and its future infrastructure commitments put its next financial disclosures among the figures to track.

For now, the evidence supports a narrower conclusion than the phrase “zero payoff” can suggest: broad gains are not yet visible in the cited aggregate measures, while some targeted uses have reported improvements. Whether those task-level results become durable, company-wide financial gains remains unresolved.

Key Questions

What did Apollo mean by “zero financial payoff”?

Apollo economist Torsten Slok used the phrase to describe aggregate trends showing rising revenue per worker and margins for seven large technology companies, but little change in those measures for the other 493 S&P 500 firms. It is an interpretation of broad data, not proof that every company outside tech has received no benefit.

Does the evidence show that AI has no productivity benefits?

No. The cited economy-wide measures have not shown a clear AI-related productivity lift, but some specific studies report gains, including more customer-support issues resolved per hour. Those results may not apply across all jobs, firms or industries.

What does Meta’s $825 billion figure include?

It refers to future leases and contractual commitments, including data centers, energy infrastructure, cloud capacity, servers and network equipment. The source material says the total also includes some Reality Labs hardware, so it should not be treated as an exact measure of AI spending or as money already paid.

Why are investors concerned about AI spending?

Companies are committing large sums to infrastructure, while broad financial returns remain difficult to identify. Meta’s reported negative free cash flow in the third quarter and rising financing costs cited in the source material have added attention to whether future returns will support the expense.

What evidence would show that AI investment is paying off?

More persuasive evidence would include sustained improvements in productivity, revenue per employee, margins or cash flow that companies can link to AI use. Results from individual workflows can help, but they do not by themselves prove gains across a whole company or sector.

Source: ThorstenMeyerAI.com