Thursday's big stock stories: What’s likely to move the market in the next trading session

Markets logged a third consecutive winning day, while the 10-year Treasury yield ended at 5%, little changed. After-hours earnings moves were mixed: Microsoft gained 3% after beating estimates, Alphabet fell about 10% after missing, and Chipotle rose 7% despite an earnings miss; neocloud stocks declined after Microsoft’s capex came in below expectations. Investors were also watching upcoming September PCE data, with consensus at +0.5% month over month and +3.7% year over year, while Jeffrey Gundlach cited a December rate hike as his base case.

Analysis

The key read-through is not simply “AI spending is slowing”; it is that the market is testing whether compute demand converts into durable, financeable contracts. A lower-than-expected Microsoft capex figure pressures neoclouds twice: weaker near-term utilization expectations and potentially tighter funding terms if lenders question collateral values or customer concentration. Microsoft may gain from capex discipline, but any sustained reduction in infrastructure orders would eventually reach equipment suppliers; Caterpillar’s data-center exposure is therefore a watch item, not a clean hedge.

The split between Microsoft’s positive reaction and Alphabet’s sharp selloff raises dispersion risk across mega-cap tech: investors are rewarding execution, not the AI label. Near term, earnings reactions and forward capex commentary matter more than broad AI sentiment. Over 1–3 months, verify cloud backlog, capacity utilization, customer concentration, and neocloud debt/refinancing terms before treating drawdowns as value. Over 6–18 months, internalized compute or improved utilization could stabilize the ecosystem; continued capex restraint would expose leveraged infrastructure models.

Rates add a cross-asset constraint. A high long-end yield and renewed hike talk make long-duration growth and weaker-credit funding more vulnerable; the upcoming PCE/jobs data can quickly reverse that pricing. High-yield ETF distribution yields alone do not establish adequate compensation for default or spread risk. Contrarian risk: large neocloud declines may already discount a severe slowdown, so crowded shorts can squeeze if Microsoft’s capex proves timing-related rather than demand destruction.

Key Decisions for Investors

• Favor quality over leveraged AI infrastructure: consider a relative underweight in CoreWeave, Nebius, and IREN versus profitable platform exposure, rather than adding to the neocloud drawdown. Reassess only after company disclosures show funded capacity, utilization, and customer commitments; a renewed capex slowdown or worsening refinancing terms would confirm the thesis.
• Do not chase the Microsoft-versus-Alphabet earnings divergence after-hours. Use the next session’s price discovery, then monitor cloud growth, backlog, and forward capex commentary over the next 1–3 months. A recovery in Alphabet’s operating outlook or evidence that Microsoft’s capex restraint is broader demand weakness would invalidate a relative-long Microsoft bias.
• Keep duration and lower-quality credit risk measured into PCE and jobless claims. Avoid using high-yield ETF yields as a standalone buy signal; a material fall in yields alongside stable credit spreads would ease the caution, while renewed Treasury-yield highs or widening high-yield spreads would argue for further de-risking.
• Treat Caterpillar’s AI-data-center linkage as a catalyst to verify, not an earnings substitute: look for order/backlog commentary and evidence of project delays before changing exposure. Broad weakness in data-center construction or order guidance cuts would falsify the near-term support thesis.