Apollo Chief Economist Torsten Slok: AI in Tech Yields Zero Financial Returns
Apollo Chief Economist Torsten Slok pointed out in a daily report that despite massive investments and widespread hype, artificial intelligence has yet to deliver any financial returns for U.S. companies outside the tech sector. Revenue per capita and profit margins have only risen among the seven tech giants, while the other 493 companies in the S&P 500 remain flat.
The report indicates that there is still no significant improvement in macro productivity data attributable to AI; mainstream companies are beginning to view token spending as a cost center that erodes profits rather than an investment that drives growth, while venture capital-backed companies continue to increase token consumption.
Slok emphasized that as long as agent-based AI and token spending remain confined to a small part of the economy, the macro effects will be very limited; the AI narrative will ultimately be determined by the 493 companies in the S&P rather than the seven giants, and there is currently no evidence that revenue growth or profit margins outside the tech sector are improving.
Data also shows that AI spending is highly concentrated, with the top 10% of companies in categories like model services and inference, new cloud, and cloud computing capturing the vast majority of the market share; about 80% of the top 1% of AI spenders are venture-backed companies, while traditional enterprises neither need the latest models updated daily nor are willing to pay high prices for these capabilities.
As a large alternative asset management company, Apollo continuously tracks this divergence through its chief economist; the report's charts compare inflation-adjusted revenue per capita, profit margins, and productivity trends, with sources including publicly available data from Bloomberg and Macrobond.
Funding flows remain concentrated in the seven tech giants and venture-backed AI consumers, while traditional companies are reducing spending on high-priced tokens to protect profit margins; the seven giants benefit from providing infrastructure and models, while non-tech companies are pressured as they treat AI as a cost rather than a revenue driver, with macro productivity yet to show diffusion effects.
The report's bottom line concludes that even if models gradually become commoditized, the key will be whether mainstream companies can convert increasingly cheaper intelligence into higher revenue growth and profit margins, for which there is currently no supporting evidence.
Source: Public Information
ABAB AI Insight
Slok has previously warned in Apollo's daily reports and public interviews that AI valuations depend on profit margin improvements outside the tech sector, but in regulated fields like healthcare, banking, and manufacturing, process transformation and data governance may extend the return cycle far beyond market expectations. If this extends from months to years, it could trigger a revaluation of valuations.
Apollo's capital path focuses on alternative asset management and macro research output, influencing institutional investors' judgments on AI capital expenditures through continuously published data-driven reports, concentrating resources on tracking the per capita metric divergence between the seven giants and the remaining companies, rather than directly investing in model training.
This is similar to the early internet bubble path where "infrastructure providers profit first, while application layers lag behind," and also resembles the early days of cloud computing where only a few software companies demonstrated efficiency gains; the current phase is a validation period following an expansion phase, with the tech seven having captured revenue while non-tech companies are still in the cost confirmation stage.
Essentially, this represents a transfer of pricing power: model and computing power suppliers (the seven giants and cloud vendors) control token pricing, while buyers reclassify spending from growth investments to cost centers, leading to funds flowing back from a broad range of companies to at least a few tech sellers until measurable revenue or profit margin improvements appear in the non-tech sector.
ABAB News · Cognitive Law
- Those selling shovels profit first, while miners are still calculating costs.
- The seven giants increase profits while 493 companies are paying the bills.
- Macro returns = diffusion to non-tech, not concentrated on sellers.