Next week brings major risks for the stock market, just as the AI trade keeps faltering

Next week brings earnings from 163 S&P 500 companies, including Alphabet, Microsoft and Amazon, alongside a Fed meeting and September PCE inflation data, as investors reassess AI spending and the tech trade. Meta fell more than 8% this week, while semiconductor stocks are down over 15% for October; the Nasdaq is set to finish the month down more than 3%. The Fed is widely expected to hold its 3.75%-4.00% rate range, while September headline PCE is forecast to rise 0.5% month over month and core inflation to reach 3.1% year over year. So far, 82% of reporting S&P 500 companies have beaten expectations, and Q3 earnings growth is projected at roughly 30% year over year.

Analysis

The key event risk is not capex in isolation; it is whether incremental AI spend is converting into durable revenue and cash flow quickly enough to defend returns on invested capital. Rising spend can support data-center and grid buildout names such as Digital Realty (DLR), Quanta Services (PWR) and GE Vernova (GEV), yet still weigh on hyperscaler multiples if payback periods lengthen. The near-term stock reaction may therefore be asymmetric: disciplined spend or evidence of monetization could relieve a crowded skepticism trade, while higher spend without clearer returns would reinforce multiple compression. Verify each company’s capex guidance, cash conversion and AI-related revenue disclosures rather than treating a headline spending number as a standalone signal.

The Fed/PCE sequence creates a separate duration risk. Sticky core inflation or firmer energy pass-through could lift yields and pressure long-duration growth even if earnings hold up; a credible patient Fed response may instead let earnings breadth matter more. Over 6–18 months, sustained data-center demand could benefit power and infrastructure suppliers, but project timing, utility rate recovery and financing costs are important constraints. The contrarian opportunity is breadth: market weakness in AI-linked hardware may be masking healthier non-tech earnings, but this is not a clean broad-market buy until rates and hyperscaler returns are clearer.

Key Decisions for Investors

• Avoid adding concentrated GOOG, MSFT or AMZN exposure immediately ahead of results. Reassess after guidance: favor the name(s) pairing controlled capex with improving cash conversion or explicit monetization evidence; treat capex acceleration without return metrics as a short-term negative.
• Consider a modest 1–3 month relative-value position: long equal-weight S&P 500 exposure (RSP) versus Nasdaq-100 exposure (QQQ), preferably after the Fed/PCE releases. This expresses breadth over AI-duration concentration; reduce or exit if core inflation eases and hyperscaler disclosures trigger a durable tech-led reversal.
• Keep PWR, GEV and DLR on a 6–18 month watchlist rather than buying the capex narrative indiscriminately. Add only if hyperscalers sustain project commitments and supplier/backlog commentary confirms execution; falsify on capex cuts, delayed projects or weaker order/backlog trends.
• Use the Fed and PCE as explicit risk controls: a hawkish repricing in yields argues for smaller growth exposure, while stable yields plus credible hyperscaler cash-return evidence would invalidate the relative underweight to QQQ.