Here are Thursday's biggest analyst calls: Nvidia, Amazon, Microsoft, Meta, Alphabet, Palantir, AbbVie & more

Wall Street calls were broadly constructive, with multiple upgrades and reiterated buy/overweight ratings across AI, software, semiconductor, and consumer names. Notable moves included upgrades to Equinix, SiteOne, Blackbaud, Teradyne, Nutrien, AbbVie, and Everus, plus new bullish coverage on Mistras and Palantir, while Meta was downgraded to neutral on intensifying AI competition. The article is mainly a compilation of analyst actions, so the impact is stock-specific rather than market-wide.

Analysis

The tape is telling us investors are rewarding businesses where AI is translating into visible monetization or operating leverage, while punishing names where AI spend is still a cost center. That creates a clear relative-value split: infrastructure and software with pricing power and utilization upside are being re-rated, while platforms facing heavier capex or marginally slower ad efficiency are at risk of multiple compression even if fundamentals remain solid.

The more interesting second-order effect is that the winners are not all the same trade. Hyperscale and cloud leaders can support a longer capex cycle, but the marginal benefit is increasingly shifting downstream to software and data-layer names that can monetize enterprise workflows faster without as much balance-sheet strain. That means the market is starting to favor “AI picks-and-shovels with revenue visibility” over “AI optionality,” especially where earnings revisions can compound over the next 2-4 quarters.

In contrast, the names tied to cyclical procurement or inventory digestion look like delayed-move stories rather than broken stories. Near-term underperformance in test equipment, smartphone exposure, or ad-platform return skepticism should persist until management teams can demonstrate a cleaner demand inflection or a more credible capex-to-ROIC bridge. This is a months-long debate, not a days-long one, so dips may continue to be bought selectively but not aggressively.

The contrarian read is that the market may be over-penalizing platform capex just as usage-based AI monetization inflects, and underestimating how quickly operating leverage can show up once model traffic becomes product behavior. At the same time, some of the recent upgrade cluster likely reflects post-earnings relief rather than fresh information, so chasing the whole basket indiscriminately is lower-quality than owning the names with clear backlog, retention, or margin expansion catalysts.

Key Decisions for Investors

• Long GOOGL / short META for 1-3 months: GOOGL has clearer evidence of AI monetization and backlog visibility, while META is more exposed to a multiple reset if capex keeps outpacing conviction on returns; target a 8-12% relative spread.
• Buy EQIX on weakness into the next 2-6 weeks: the trade is not about immediate earnings surprise, but about re-rating as AI workloads sustain occupancy and pricing power; use a tight stop if guidance fails to confirm utilization acceleration.
• Go long PLTR versus QCOM for a 3-6 month horizon: PLTR is an earnings-revision and multiple-expansion story tied to AI software adoption, while QCOM remains tied to a slower handset digestion cycle; asymmetry favors the former if enterprise spend stays resilient.
• Consider long MSFT calls financed by short-dated upside sales on a pullback: the market is still underappreciating the durability of usage-based monetization; risk is capex fatigue, but the payoff improves if next-quarter cloud commentary remains constructive.
• Avoid blanket chasing of the AI basket; instead pair long NVDA against short a cyclical semiconductor test-exposure name like TER if you want a cleaner AI leader vs procurement-lumpiness expression over the next quarter.