# Outside Big Tech, the AI payoff is still missing

> Apollo's Torsten Slok published charts on 24 October showing that revenue per worker, profit margins and productivity remain flat outside the Magnificent Seven. More than half of American businesses now pay for AI. Findings from a Stanford dashboard on employment, the Federal Reserve's Beige Book and last week's fall in the share prices of data-centre suppliers are consistent with his account. Slok argues that the payoff will depend on companies outside Big Tech, where he sees no sign of it yet.

- URL: https://ai-news-daily.xyz/posts/outside-big-tech-the-ai-payoff-is-still-missing/
- Published: 2026-10-25
- Tags: business, research
- Author: AI-generated (this text was generated by AI; see https://ai-news-daily.xyz/about/#ai-disclosure)
- Publisher: [AI News Daily](https://ai-news-daily.xyz/)

The meeting described here never happened, and the people in it are fictional illustrations. Their decision illustrates the behaviour that Torsten Slok documented on 24 October. A sales engineer for an AI platform visits a regional food distributor to show its finance team what the newest model can do with their invoices and call transcripts. They watch the demonstration through. They already use a cheaper, year-old model on a lower subscription tier. With that model, they finish processing the queue most days by lunch. When the engineer names the price of the upgrade, the chief financial officer sets down her pen. The team keeps the same number of subscriptions.

Slok is Apollo's chief economist. In his Daily Spark note that Saturday, he reached a firm conclusion: "Despite massive investments and widespread hype around artificial intelligence, the technology has delivered zero financial payoff to corporate America outside of tech." He based that conclusion on charts that compare two groups of companies. One group is the Magnificent Seven, the large technology firms that build and sell the models. The other comprises the remaining companies, which Slok calls the S&P 493.

## The charts compare two parts of the market

The first two charts track revenue per worker and profit margins. According to Slok's reading, both measures are rising for the Mag 7 and both are flat for the S&P 493. He uses Bloomberg and Macrobond data adjusted for inflation to the fourth quarter of 2019. The third chart is the main basis for his broader claim. It plots total factor productivity, the part of output growth that remains after accounting for labour and capital, and shows no acceleration. In Slok's words, "there are still no signs of AI boosting productivity growth."

Apollo does not produce that productivity series. The Federal Reserve Bank of San Francisco maintains it, using a method that John Fernald described in a 2014 working paper. The measure adjusts total factor productivity for changes in how intensively firms use their workers and equipment through the economic cycle. By removing those fluctuations, the adjustment gives a closer estimate of genuine technological change. That makes it a useful measure for assessing a general-purpose technology. So far, it shows no acceleration.

Slok links those results to how companies spend. With no improvement in their financial results, he writes, "mainstream corporate America is pulling back and starting to view token spending not as a source of growth but as a cost center that weighs on margins." The fictional food distributor's decision not to upgrade illustrates that argument. In Slok's account, mainstream firms do not need the latest model and are unwilling to pay for it.

## Widespread adoption has not produced a visible payoff

The lack of a visible payoff is more striking because the tools are widely used. Economist Ara Kharazian at Ramp's economics lab maintains the Ramp AI Index, a monthly measure of business adoption of AI. Its September reading put adoption at 55.9% of American companies. More than half of the country's firms are paying for AI, yet the financial benefit is still not visible in these measures.

Slok raises a second concern about the companies that spend the most. In a companion note two days earlier, he cited Ramp data showing that 80% of the top 1% of AI spenders are venture-backed, compared with only 13% of the median spender. The largest buyers of computing capacity are paying with investors' capital rather than their own earnings. He argued that a slow payoff would put pressure on venture returns. If investors then reduced their funding, demand for computing capacity would weaken too. Observers have identified similar circular dependencies in deals between chipmakers, cloud providers and model labs. Slok's figures suggest that those dependencies also extend to the companies that pay to run the models.

## Evidence from employment and the stock market

Other findings this month were consistent with Slok's account, although they came from different sources and measured different things.

The first concerns employment. Stanford's Digital Economy Lab maintains public dashboards that track AI's effects on employment. The dashboards launched in June and are updated monthly. In a 21 October write-up, the university reported falling employment among workers aged 22 to 25 in occupations that are heavily exposed to AI, including software development and customer service. The recent decline was 4.4%. The dashboards grew out of a paper by Erik Brynjolfsson, Bharat Chandar and Ruyu Chen, who described these young workers as canaries in the coal mine.

The Federal Reserve reported related findings from employers. Its October Beige Book recorded that "Five Districts reported that some firms had either reduced staffing levels (through attrition), delayed hiring, or anticipated less job creation after increasing productivity by investing in AI". This was the first edition in which the survey's recurring sentence about AI described reductions in staffing rather than the introduction of tools.

The change is sharper inside firms that do report a payoff. In a 16 October report, Anthropic economists Nathan Wilmers, Regina Catipon, Maxim Massenkoff and Peter McCrory found that the proportion of junior technical hires at the company "fell even faster, from 27% to 9%," and that by August its agents "worked 6 times as many hours as their human handlers." This illustrates how benefits can be concentrated in a technology firm, with part of the consequence appearing as jobs that are never created.

Stock prices provide another comparison. On 23 October, Yahoo Finance's Jared Blikre calculated the changes since the S&P 500's record close on 6 October. He found that "the companies building and equipping AI data centers are down a median 23%," while "the Big Tech companies writing the checks are down just 5%." The suppliers' share prices fell four times as much as those of their customers, even though suppliers had record backlogs of orders. Investors appeared to be betting that spending would slow before the returns became visible.

## Whether the payoff will arrive remains uncertain

Slok leaves open the question of whether this difference will narrow. Models are becoming cheaper and more alike, but he asks whether mainstream firms can use that cheaper intelligence to increase revenue and profit margins. His answer is cautious. "AI may be a great calculator," he writes, "but in business, intelligence is mostly judgment." Judgment depends on context and on the person making it. A model, he adds, "does not understand accountability, trust, culture, unwritten rules, ethical trade-offs or fairness," all of which affect whether a customer stays or a deal is completed.

That is Slok's argument, not a measured result. The charts show that the measures remain flat for the S&P 493; they do not establish that AI can never improve them. They show only that no improvement is visible yet. A single measure of productivity adjusted for utilization can miss real gains that take time to accumulate. Similarly, a figure for adoption counts subscriptions, not how much people use them.

Back in the fictional food distributor's office, the finance team makes its usual calculation. The year-old model is good enough. The newer one costs more than the value of the work that it would save, and the team must complete the quarter's accounts either way. Meanwhile, revenue per worker continues to rise within the Mag 7 in Slok's chart. That difference is the central issue. The same measures over the next few quarters will help show whether other firms are waiting for a delayed payoff or whether that payoff continues to be absent.

## References

- Blikre, J. (2026). *The AI sell-off is crushing stocks of companies leading the data center build-out*. Yahoo Finance. https://finance.yahoo.com/markets/article/the-ai-sell-off-is-crushing-stocks-of-companies-leading-the-data-center-build-out-175206076.html
- Brynjolfsson, E., Chandar, B., & Chen, R. (2026). *Canaries in the Coal Mine? Six Facts about the Recent Employment Effects of Artificial Intelligence*. Stanford Digital Economy Lab. https://digitaleconomy.stanford.edu/publications/canaries-in-the-coal-mine/
- Federal Reserve System (2026). *Beige Book, October 2026: National Summary*. https://www.federalreserve.gov/monetarypolicy/beigebook202610-summary.htm
- Fernald, J. (2014). *A Quarterly, Utilization-Adjusted Series on Total Factor Productivity*. Federal Reserve Bank of San Francisco. https://www.frbsf.org/research-and-insights/data-and-indicators/total-factor-productivity-tfp/
- Hadhazy, A. (2026). *Tracking AI's Real-Time Footprint on the Labor Market*. Stanford HAI. https://hai.stanford.edu/news/tracking-ais-real-time-footprint-on-the-labor-market
- Kharazian, A. (2026). *Ramp AI Index*. Ramp Economics Lab. https://ramp.com/data
- Slok, T. (2026). *Circularity in AI Shows Up in Token Spending Too*. Apollo, The Daily Spark. https://www.apollo.com/wealth/insights-news/insights/daily-spark/circularity-in-ai-shows-up-in-token-spending-too
- Slok, T. (2026). *No Signs of AI Payoff Outside the Tech Sector*. Apollo, The Daily Spark. https://www.apollo.com/wealth/insights-news/insights/daily-spark/no-signs-of-ai-payoff-outside-the-tech-sector
- Wilmers, N., Catipon, R., Massenkoff, M., & McCrory, P. (2026). *How employment is changing at Anthropic*. Anthropic. https://www.anthropic.com/research/employment-at-anthropic