AI Spending Is Lifting Big Tech Margins. Everyone Else Is Waiting.
Apollo's chief economist finds AI capex is boosting margins for the Mag Seven to ~25%, but the other 493 S&P 500 companies are stuck at ~10%. Here's why it matters.
Apollo chief economist Torsten Sløk published a report finding that the AI capital spending surge is improving profit margins almost exclusively for the companies selling AI, not the companies buying it. The Magnificent Seven tech group now posts net margins of around 25%, up from about 15% in late 2023. The remaining 493 companies in the S&P 500 are stuck near 10%, roughly where they have been for three years. Sløk warns that investors are expecting proof of returns by 2028, and a slow payoff could push markets into correction territory.
What happened
| Data point | Detail |
|---|---|
| Mag Seven net margin (current) | ~25% |
| Mag Seven net margin (late 2023) | ~15% |
| S&P 493 net margin (current) | ~10% |
| S&P 493 margin trend (3 years) | Flat |
| AI ChatGPT launch (baseline date) | End of 2022 |
| Investor ROI expectation deadline | Around 2028 |
Torsten Sløk, chief economist at Apollo, published a report noting a clear split in who is actually profiting from the AI spending wave. Since OpenAI launched ChatGPT at the end of 2022, net profit margins for the technology and communications services sectors have climbed. Every other major sector, including healthcare, consumer staples, energy, and materials, is either flat or down.
Put simply, Nvidia and the major cloud platforms are the sellers of AI infrastructure. Their margins are rising. The thousands of companies paying for that infrastructure are the buyers, and their margins have not moved.
“The AI capex boom is so far only showing up in the sellers’ margins, not the buyers’,” Sløk wrote.
Why it matters
The stock market is priced for a very optimistic AI outcome. Major indexes are near record highs, driven heavily by enthusiasm around AI. That optimism is concentrated in a small number of names. If the broader market does not see returns, the risk is not just a sector-level miss.
“With so much riding on so few names, a slower payoff wouldn’t just be a sector problem, it would risk tipping the economy into recession and the S&P 500 into a correction,” Sløk said.
Sløk noted in a separate early-July report that many investors expect meaningful AI returns to show up by 2028. That gives non-tech companies roughly two years to demonstrate that their AI spending produces real earnings gains. Rob Almeida, global investment strategist at MFS Investment Management, framed the stakes plainly: “This cycle is dependent on AI being able to monetize at the rate the market thinks it will.”
Sløk told Barron’s he does expect margins outside tech and telecom to eventually expand. The question is speed, not direction. A slow rollout means years where the market bet sits exposed.
What does this mean for businesses spending on AI tools?
For a typical business owner investing in AI subscriptions, custom models, or workflow automation today, the macro data matches what many are experiencing on the ground: costs are real, savings are partial, and the efficiency gains are still maturing. The margin picture at the S&P 493 level is essentially a large-scale version of that same experience.
The gap also highlights where the value currently lives. Companies that build or sell AI capabilities, including the platforms running large language models like ChatGPT and Claude, are capturing the value first. Buyers of those tools are paying a premium and waiting for the compounding returns.
If you are weighing AI investments for your own business, our AI integration services focus on deployments that produce measurable output changes rather than subscription costs with unclear returns. The margin data from Apollo is a useful reminder to track what AI spending actually changes in your cost structure before expanding it.
Our take
Sløk’s numbers confirm what we have suspected for a while: the productivity story is real but early, and right now the infrastructure layer is capturing most of the value. Businesses outside tech are essentially funding the margin expansion of the companies they buy from.
That does not mean AI tools are not worth using. It means the bar for proof should be higher than “we deployed it.” For any AI project, you need a baseline, a measurable outcome, and a clear cost comparison before and after. Otherwise you are contributing to someone else’s 25% margin.
The 2028 deadline framing is useful for planning. If broad ROI does not materialise by then, expect significant repricing of AI-exposed stocks, which will shift vendor pricing, investment appetite, and tool availability. Watch for that window. We track AI developments and market signals regularly if you want to stay ahead of those shifts.
What to do about it
- Audit your current AI tool spending and map each cost to a specific business process it is supposed to improve.
- Set a measurable baseline now: revenue per employee, support ticket volume, content production time, or whatever the tool is meant to affect.
- Review that metric at 90 days and 180 days. If the line has not moved, cut or renegotiate before the subscription auto-renews.
- Focus new AI investment on workflow automation with clear input/output ratios rather than broad platform licenses with vague productivity promises.
The margin data is a signal, not a verdict. Track your own numbers the way Sløk tracks the S&P 493.
Frequently asked questions
Is AI actually improving company profit margins?
According to Apollo chief economist Torsten Sløk, AI has improved margins for tech and communications companies (the Magnificent Seven sit around 25% net margins), but the other 493 S&P 500 companies have seen flat or declining margins since ChatGPT launched in late 2022.
When will businesses outside tech see returns from AI investment?
Sløk noted in an early-July 2026 report that many investors expect tangible AI returns to appear by around 2028, giving non-tech companies roughly two years to demonstrate earnings gains from their AI spending.
What are the Magnificent Seven's current profit margins?
According to Sløk's data, the Magnificent Seven tech companies have net margins of around 25%, up from about 15% in late 2023.
What happens if AI doesn't deliver returns by 2028?
Sløk warned that a slow payoff, given how concentrated the market is in AI names, could tip the economy into recession and push the S&P 500 into a correction.
