Apollo Analysis Questions AI’s Financial Payoff Beyond Big Tech

Apollo Global Management’s newest economic study is asking a hard question for companies in America: are companies that are not in the technology world seeing clear financial gains from their increasing spend on artificial intelligence?

In a research note dated October 24 Apollo partner and chief economist Torsten Slok said that even though many companies are using AI and spending a lot there is still little proof that AI has made revenue per worker corporate profit margins or overall productivity better across much of the United States economy. The study compares the performance of the seven technology firms, known as the Magnificent Seven with the other 493 firms in the S&P 500.

Apollo’s findings do not prove that every company outside the technology sector has failed to gain from AI. Instead Apollo shows a gap between the money results of technology companies and the rest of the corporate world raising questions about how fast AI spend will turn into clear business results.

Revenue and Profitability Show a Divide

One comparison in Apollo’s study is revenue per worker. Apollo’s charts show that this number has gone up among the Magnificent Seven. Has stayed about the same for the other 493 firms in the S&P 500. Apollo’s revenue numbers are adjusted for inflation to 2019 levels.

Apollo also notes a gap in corporate profit margins. Apollo says margins have grown among the technology companies but have stayed about the same across the rest of the index.

These numbers matter because companies usually need more than new technology to justify big spending. If AI tools help workers finish tasks faster repeated work or make better decisions companies would hope to see clear results such as higher output per worker stronger revenue, lower operating costs or better margins.

Financial performance is affected by many things beyond AI. Demand, pricing, wages, interest rates, competition and changes in business strategy can all affect revenue and profit. Apollo’s comparison shows a lack of wide financial improvement but it does not prove on its own that AI caused the gap.

Productivity Gains Remain Difficult to Identify

Apollo’s note looks at economy-wide productivity data and finds no clear proof that artificial intelligence has yet caused a big increase in productivity growth.

Productivity matters because it shows how well an economy turns other inputs into goods and services. A technology that many people use can have an economic effect if it lets workers and businesses make more with the same resources.

The effects of big technologies do not always show up right away in national data. Companies may need time to reorganise work, train workers, link systems with old software and set up reliable processes. Some benefits may first appear in tasks or industries not all across the economy.

In January 2026 Apollo’s own research said that academic studies on AI’s productivity impact were still unclear while pointing to gains in areas such as DevOps software, robotic process automation and content management systems. Apollo’s newer analysis puts focus on the lack of clear economy-wide financial payoff so far.

These two studies show the difference between gains in specific applications and proof of a clear change across the whole economy.

The Challenge of Measuring AI Returns

The question is getting more important as companies grow their use of AI assistants, automated workflows and AI-powered enterprise software.

In an analysis published in September Slok said that about 69% of S&P 500 companies were actively using AI according to second‑quarter 2026 earnings disclosures.. Only 29% reported a measured AI result and about 2% tracked a relevant metric over time according to Forbes figures.

The numbers show that using AI and showing value are different things. A company can bring in AI tools without proving that they cut costs raise output or improve customer results to justify the spend.

Measuring returns can also be hard. Savings may be spread across departments while costs can include subscriptions, computing use, data preparation, cybersecurity, worker training and linking with existing systems. Some benefits, like research or better customer service may be hard to turn into short‑term profit.

For leaders the real challenge is to set a clear baseline track results the same way and separate real improvements, from changes that would have happened anyway.

AI Spending Faces Greater Scrutiny

Apollo says that some mainstream businesses are starting to see spending on AI services and model usage as a cost centre of an automatic source of growth. This could make a challenge for companies that sell AI infrastructure and software if customers get more selective about how much they spend and which tools they use.

The issue is not simply whether businesses will stop using AI. Companies may keep adopting tools that deliver benefits while cutting back on services that cannot show enough value. They may also choose lower‑cost models or narrower applications than the most expensive systems.

That would shift the market toward results. Vendors could feel pressure to demonstrate how their products improve productivity cut costs or generate revenue instead of just relying on general hopes about future AI adoption.

Investors are also watching how large capital spending relates to earnings. Technology companies have benefited from demand for chips, data centres, cloud computing and AI services.. The long‑term financial case depends partly on whether businesses that buy those services can create enough value to keep spending.

What Businesses and Investors Should Watch

The next phase of AI adoption might depend on whether companies can turn deployments into repeatable improvements in everyday operations.

Businesses will need to track outcomes such as processing time, error rates, customer satisfaction, employee output and total cost of delivering a service. The relevant measures will differ by industry and by the task being automated. A useful result in customer support for example may not be measured the way as a result in manufacturing or financial analysis.

Investors meanwhile will search for evidence that AI spending is adding to revenue growth cost efficiency or stronger margins. Corporate earnings reports, capital expenditure plans and disclosures about results could help show whether that transition is happening.

Apollo’s analysis warns against assuming that widespread technology adoption automatically leads to economy‑wide productivity gains. It does not say that AI has no value nor does it rule out benefits as systems improve and businesses learn how to deploy them effectively.

For now the central question stays unresolved: can companies, outside the technology sector turn AI investment into financial gains? The answer will depend on business results, not just the amount of money spent or the number of tools adopted.