AI-related companies to drive most third-quarter US earnings gains

Analysts expect S&P 500 third-quarter earnings to rise about 31% year over year, with roughly two-thirds of the increase attributed to technology and AI-heavyweights; energy earnings are forecast to climb about 115%. The outlook is tempered by concerns that earnings growth is nearing its peak and estimate-revision momentum is cooling: semiconductor earnings growth is projected at about 136%, down from 158% in Q2. Investors will also watch the effect of higher interest rates on corporate profits, particularly for heavy borrowers such as utilities.

Analysis

The key risk is not an outright earnings contraction; it is a narrowing of the earnings engine. With a large share of expected growth concentrated in AI-linked companies, a market-level “beat” can conceal weak breadth and leave index multiples unusually sensitive to any capex pause or reduction in forward estimates. That creates asymmetric event risk: results can meet consensus yet disappoint if management commentary fails to raise the next-quarter hurdle.

The second-order beneficiaries remain memory and power infrastructure, but confirmation matters more than headlines. Micron’s customer commitments support demand visibility, not necessarily realized revenue, pricing or margins; monitor conversion into shipments and pricing commentary. Constellation Energy’s power agreement validates the importance of reliable electricity for data centers, but without contract economics it does not establish near-term earnings accretion. Conversely, higher yields pressure capital-intensive and heavily indebted businesses, while elevated oil may support energy earnings but also feed inflation and rates—potentially offsetting the benefit to the broader index.

Over the next few weeks, bank results and management guidance should test whether earnings strength extends beyond AI and commodity-linked names. Over 6–18 months, the pivotal variable is whether AI investment produces monetizable returns fast enough to sustain capex. Consensus may be underpricing the breadth problem even as it recognizes that semiconductor growth is slowing; this is a reason to hedge index concentration, not to declare the AI buildout over.

Key Decisions for Investors

• Avoid adding broad S&P 500 exposure solely on headline earnings growth. Consider a defined-risk 1–3 month S&P 500 put spread into the reporting window if option pricing is reasonable; reduce or remove it if earnings revisions broaden beyond AI and energy or index breadth improves.
• Keep Micron (MU) as a selective, event-driven long rather than a sector-wide AI proxy. Add only if results confirm customer commitments are converting into shipments and pricing; cut the thesis if guidance or memory pricing weakens. The main risk is that commitments defer rather than secure profitable near-term sales.
• Treat Constellation Energy (CEG) as a watch item, not an automatic buy on the power-deal headline. Verify contract duration, pricing, capital requirements and expected earnings contribution; higher yields could offset any demand-driven upside.
• Use the season to test earnings breadth: track revisions and guidance from non-AI sectors alongside JPMorgan (JPM) and Goldman Sachs (GS). A clear deterioration in forward estimates, or further yield increases, would strengthen the hedge; broad-based upward revisions would falsify the concentrated-earnings downside.