Stocks making the biggest moves midday: Amazon, Marvell, PayPal, Salesforce & more

Workday shares jumped 6% after Q2 results beat estimates ($2.75 adj. EPS and $2.65B revenue vs. $2.61 and $2.64B). In contrast, PayPal sank nearly 16% after reports that Advent and Stripe decided not to pursue a buyout, while Marvell fell 10% on softer Q1/Q2 outlook. Crypto-related names dropped with bitcoin around $77.6K, as Affirm rose 5% on fiscal Q4 revenue of $1.17B beating the $1.11B estimate.

Analysis

The tape is rewarding software names where AI is already showing up in current bookings, retention, or deal conversion, and punishing anything where AI still looks like an upstream cost center. That favors platform/app names like WDAY, CRM, NOW, ESTC, and AMZN over semis where the market is suddenly demanding clean margin expansion, not just AI demand visibility; MRVL is the clearest example of a stock that can miss the bar even when the business is still growing. The second-order effect is a widening valuation gap inside tech: installed-base software can pass through AI capex into pricing, while component suppliers get no credit unless gross margin inflects immediately.

PYPL is the most interesting loser because the absence of a strategic bid removes the implicit downside floor. Without that optionality, the equity has to trade on execution against faster checkout and BNPL alternatives, which means the stock can drift lower over weeks even if near-term operating metrics are merely stable. AFRM’s strength suggests capital is still rotating toward consumer-finance models with clearer product-market fit, pressuring legacy wallet/payment rails and merchant ecosystems more than consensus implies.

Contrarian view: the market is treating every AI-positive print as durable, but much of the upside is still sentiment-driven and can reverse if next-quarter margin or guidance disappoints. The reversal triggers are simple: MRVL needs margin stabilization, PYPL needs a credible strategic catalyst, and BTC needs to reclaim momentum for crypto beta to stop leaking. Until then, this is a stock-pickers’ tape rather than a broad risk-on signal.

Key Decisions for Investors

• Long WDAY / CRM / NOW basket vs short MRVL for 1-3 months: own the names where AI is monetizing inside the P&L and fade the supplier where margin math is still under scrutiny. Risk/reward is attractive if software keeps re-accelerating and semis remain range-bound; cut if MRVL regains margin confidence or if software guides only maintain, rather than raise, estimates.
• Short PYPL on any bounce over the next 2-6 weeks. The strategic-buyer floor is gone, so the stock now trades as a competitive-displacement story rather than an M&A optionality story; downside can persist until there is evidence of take-rate stabilization. Falsify if management announces a meaningful capital-return step-up or a credible transaction emerges.
• Pair long AFRM vs short PYPL into the holiday/consumer-spend setup, 1-3 months. AFRM has the cleaner growth narrative while PYPL carries the burden of defending share in checkout; this pair isolates relative share shift rather than absolute consumer demand. Risk is higher loss severity in AFRM if credit performance deteriorates or funding costs widen.
• Tactical long AMZN on pullbacks over 1-3 months. The upside is not just retail; if AI meaningfully improves search, conversion, and ad monetization, estimates can grind higher without heroic assumptions. Falsify if incremental AI usage proves to be survey noise rather than measurable basket-size or margin lift.
• Avoid chasing NVDA strength until the next catalyst window; use it more as a hedge against over-owned AI software longs than as a standalone momentum add. The near-term risk is mean reversion after a sharp move, while the structural bull case remains intact unless datacenter demand or pricing power visibly stalls.