Stock market today: Dow, S&P 500 and Nasdaq 100 futures pullback after all-time highs, Tesla slips after-hours

US stocks fell Thursday, with the S&P 500 down about 0.4% to 0.5%, the Dow off about 0.4% to 0.5%, and the Nasdaq down 0.9% to more than 1% as Iran peace talks stalled and oil surged. Brent crude moved back above $105 per barrel and WTI topped $96, while software stocks plunged after ServiceNow dropped 17.75% and IBM slid over 8% to 10% on mixed results and AI disruption concerns. The broader software ETF IGV fell 5%, overshadowing strength in semiconductors and select AI infrastructure names.

Analysis

The market is starting to price a regime shift from “AI software optimism” to “AI infrastructure scarcity.” The clearest relative winners are the picks-and-shovels names that monetize power, grid load, and data-center capex with less model-specific risk: electrical equipment, power generation, and select industrial automation should keep outperforming even if semis pause. That makes the recent strength in names tied to electrification and compute load more durable than the software selloff, because the former is supported by multi-year capacity buildouts while the latter is being reassessed on customer ROI and AI substitution risk.

The software weakness is broader than one bad print: investors are punishing anything with near-term revenue sensitivity, long sales cycles, or exposed enterprise budgets. That argues the next leg down, if oil stays elevated and rates firm, is in high-multiple software with the weakest cash conversion and greatest reliance on large deal closure timing. Conversely, the blowout in analog and mature chip names suggests the semis trade is rotating from pure AI scarcity to a broader industrial/data-center demand validation — a healthier setup, but also one that can get crowded fast if macro risk stays elevated.

Oil above $100 is not just an energy trade; it is a margin-tax on transportation, chemicals, and discretionary spending, with the most immediate second-order effect being pressure on companies already missing on guidance. The real risk is that higher fuel costs and geopolitical uncertainty slow enterprise buying decisions right as software vendors are trying to defend premium multiples. If diplomatic headlines improve, the oil bid can unwind quickly, but the software de-rating may persist because the market is now questioning the durability of next-year growth assumptions rather than just this quarter’s print.

The contrarian angle is that this may be a better buy-the-dip in infrastructure than a sell-the-rally in software. The market is treating every negative software print as AI displacement, but some of the damage is likely just timing noise and margin digestion; meanwhile, the infrastructure beneficiaries have run so far that any disappointment in grid, power, or capex cadence could trigger a sharp mean reversion. In the near term, the highest-probability setup is not broad tech beta — it is dispersion within tech, long physical capacity and short software monetization risk.

Key Decisions for Investors

• Long GEV / ETN / VRT basket versus short IGV for 2-6 weeks: infrastructure beneficiaries have clearer earnings durability, while software faces multiple compression if deal delays and AI substitution fears persist.
• Short NOW or CRM into any bounce over the next 1-3 sessions; use tight risk controls. These names are most exposed to the market repricing enterprise software growth as timing-driven rather than secular.
• Pair long TXN or ADI against short SNOW or ADBE for a 1-2 month relative-value trade: mature semiconductor/analog demand is being validated by end-market restocking, while software is absorbing a higher discount rate and tougher AI narrative.
• Add tactical long XLE or call spreads only if Brent holds above $100 for another 3-5 trading days; otherwise treat the oil spike as headline-driven and prone to reversal on diplomatic headlines.
• For event-driven traders, consider a call spread in INTC into the print with defined downside; the stock is at a long-term technical ceiling, so upside can be meaningful if guidance confirms CPU/data-center demand, but earnings risk is binary.