Apollo says AI has delivered zero financial payoff outside tech

Apollo Global Management Chief Economist Torsten Sløk said AI has delivered no financial payoff to corporate America outside the technology sector, despite substantial investment and hype. His data show revenue per worker rising in tech but flat across the rest of the economy, with no evidence of margin expansion elsewhere.

Analysis

The key market implication is a potential split between selling the tools and buying the productivity. AI infrastructure vendors can book revenue as customers invest, while corporate users may not yet capture savings—or may pass them to customers through lower prices. That supports near-term supplier demand but raises the risk that user-side capex eventually faces tougher return hurdles. It is not evidence that AI has no value: revenue per worker and reported margins can miss quality improvements, avoided hiring, or benefits that take time to appear. They can also be distorted by workforce mix and layoffs.

Over the next 1–3 months, earnings calls are the useful test: look for quantified labor-cost savings, incremental revenue, and capex payback rather than deployment counts or management anecdotes. Over 6–18 months, absent those disclosures, investors may discount AI-related spending by non-tech firms and reassess infrastructure demand. The main reversal would be credible evidence of recurring productivity gains translating into operating margins or revenue growth outside technology. No broad directional trade follows from this single data point; valuation, spending exposure, and adoption vary substantially across firms.

Key Decisions for Investors

• Do not treat AI adoption announcements as proof of customer ROI. In upcoming earnings, track realized labor-cost savings, incremental revenue, and capex payback; distinguish quantified results from pilots and usage metrics.
• Avoid adding broad, valuation-insensitive AI exposure solely on the assumption that corporate adoption guarantees durable infrastructure demand. Reassess supplier exposure if customer capex plans weaken or deployment fails to convert into recurring workloads.
• Use a relative-value watchlist rather than a blanket short: compare AI infrastructure suppliers with non-tech adopters whose spending rises without margin or productivity guidance. Consider a pair only after checking valuation, order/backlog sensitivity, and the companies’ actual AI exposure.
• Falsification trigger: repeated, measurable margin improvement or revenue-per-worker acceleration across non-tech sectors would undermine the delayed-ROI thesis; continued spending growth without such evidence would strengthen the case for more selective positioning.