AI Payoff Outside Big Tech Still Missing, Apollo's Slok Says: What It Means for Marketing ROI

Apollo chief economist Torsten Slok said on October 24, 2026 that AI has delivered no measurable financial payoff to US companies outside tech. In his Daily Spark note, revenue per worker and profit margins are rising for the Magnificent 7 but flat for the other 493 S&P 500 companies. If those buyers don't see returns, AI valuations look exposed. For marketing teams, the message is simple: prove AI ROI in numbers or expect budget cuts.

What exactly did Torsten Slok say on October 24?

Slok argued that, despite heavy spending, AI has produced zero financial payoff for corporate America outside tech. His October 24 Daily Spark rests on five charts built from Bloomberg, Macrobond, San Francisco Fed and Ramp data:

• Inflation-adjusted revenue per worker and profit margins are rising for the Mag 7 and flat for the S&P 493.
• Productivity data, including the San Francisco Fed's total factor productivity series, shows no AI boost.
• Ramp data shows token spending concentrated among venture-backed firms. Mainstream companies, he wrote, don't need the newest model and won't pay for it.

He added that mainstream firms increasingly treat token spending as a cost center. His conclusion: "the AI story will be decided in the S&P 493, not the Mag 7."

A note on dates: the headline often attached to this story, "The Buyers of AI Are Still Waiting for the Payoff," belongs to an earlier Slok note dated August 9, 2026.

How does this build on his earlier warnings?

The October note is the latest in a series arguing that the payoff outside tech is slow:

1. June 30: Slok argued the ROI runway could be long in regulated sectors such as health care, banking and energy.
2. July 6: In Fortune, he warned of a possible painful market repricing and said, "The key issue is the length of the ROI runway outside the tech sector." Fortune cited Bloomberg and Macrobond data showing Mag 7 margins rising from about 15% to about 25% between Q1 2023 and Q1 2026, while the S&P 493 stayed near 10%.
3. August 9: A sector breakdown found health care margins roughly halved since 2015 and AI gains "only showing up in the sellers' margins, not the buyers'."
4. September: Forbes reported Slok's Q2 earnings analysis: about 69% of S&P 500 companies actively use AI and 29% reported a quantified result, but only about 2% tracked an AI metric over time.

Exact Mag 7 margin levels vary by source and chart reading, but the direction is consistent across his notes.

Why would this matter for markets?

AI vendor valuations assume their customers will profit from AI and keep spending, so a long delay in payoff is a repricing risk. In Slok's framing, if non-tech buyers can't show returns, they slow purchases, and the revenue growth priced into Big Tech looks too optimistic.

Earnings data shows how concentrated the gains are. LSEG estimates cited by Yahoo Finance projected Q3 2026 S&P 500 earnings growth of about 31%, with tech and AI-related companies expected to supply roughly two-thirds of it. Energy, lifted by oil prices rather than technology, was another big contributor.

What do critics and other economists say?

Not everyone reads flat aggregate margins as proof that AI isn't working. The main counterpoints:

• The lag may be normal. Goldman Sachs economist Elsie Peng, reported by Fortune, modeled a J-curve with significant gains only around year eight, echoing how slowly PCs showed up in productivity data. Goldman still expects AI to meaningfully lift productivity over the next decade.
• Averages hide winners. A flat S&P 493 line can mask individual gains. Fortune cited a Ricoh claims project that cost more than expected but eventually tripled the division's productivity.
• Correlation isn't causation. The Mag 7's margin gains may also reflect pricing power, cost cuts and cloud scale that predate generative AI, so the charts don't isolate AI's effect.
• Slok himself expects the gains to come eventually. His argument is about timing and whether markets have priced in a fast payoff, not a claim that AI is worthless.

What does this mean for marketing teams?

Marketing is one of the clearest tests of whether the S&P 493 can turn AI spending into revenue, and measurement is the weak link. A TransUnion study of 100 senior US marketing and technology leaders, released August 5, found:

• 65% judge AI success mainly by time and cost savings.
• Fewer than half use marketing mix modeling, multi-touch attribution or incrementality testing.
• 70% report cross-channel blind spots.
• 89% expect AI marketing investment to rise anyway.

As TransUnion put it, "AI isn't a shortcut around data challenges." Time saved isn't margin gained unless it shows up as more pipeline, lower acquisition cost or freed-up budget.

What should marketing teams do now?

Treat every AI line item as an investment that has to earn its place in the budget. When CFOs start reading notes like Slok's, budgets built on hours saved are the first to be questioned.

1. Define the outcome before the tool. Tie each AI use case to one revenue or cost metric, such as cost per qualified lead, conversion rate or content cost per asset, and record a baseline first.
2. Use controlled tests. Holdout groups and incrementality tests show what AI actually changed. Before-and-after anecdotes don't.
3. Reallocate, don't just add. When an AI workflow proves out, move budget from the channels or tasks it replaces. MarkGrid's Budget Optimization module is built to model those shifts across channels, so savings show up in the plan instead of disappearing into busy work.
4. Track AI visibility as its own channel. As buyers research through AI assistants, brand presence in model answers is a measurable outcome. Model Share tracks how often and how favorably your brand appears in AI-generated answers, which gives the AI line item a metric beyond efficiency.
5. Report on a set cadence. Only about 2% of S&P 500 firms track an AI metric over time. Quarterly reporting on the same metrics puts a team ahead of nearly all of them.

For more on prioritizing, see our guide to AI marketing trends brand teams should watch. Teams that want to test this approach can start with MarkGrid's Go plan.

FAQ

What is the S&P 493?

It's the S&P 500 minus the Magnificent 7 (Apple, Microsoft, Alphabet, Amazon, Meta, Nvidia and Tesla). Slok uses it as a proxy for the AI-buying economy outside Big Tech.

Is "The Buyers of AI Are Still Waiting for the Payoff" the same note?

No. That title belongs to Slok's August 9, 2026 Daily Spark on sector margins. The October 24 note is titled "No Signs of AI Payoff Outside the Tech Sector."

Does this mean AI doesn't work?

Not necessarily. Goldman Sachs and others argue that payoffs from new technology lag by years, and aggregate data can hide individual companies that are gaining.

How can marketers prove AI ROI?

Set baselines, run holdout or incrementality tests, tie each use case to a revenue or cost metric, and move budget when results hold up.

Why does this matter for marketing budgets?

If CFOs see AI as a cost center, marketing AI spending without measurable outcomes is likely to be cut first.

Last updated October 25, 2026.