Here are Friday's biggest analyst calls: Nvidia, SpaceX, Amazon, Intel, Rivian, Coinbase, Reddit, GE Vernova & more

Friday’s Wall Street calls were mixed: Jefferies upgraded Rivian to Buy, while KeyBanc downgraded Reddit to Sector Weight on slowing search traffic and uncertainty around 2027–28 revenue and EBITDA. Morgan Stanley cut Amazon’s price target to $335 from $350 and Coinbase’s to $190 from $210, while Rosenblatt upgraded Monolithic Power Systems after a strong beat-and-raise quarter. Analysts also cited a delayed Intel turnaround, weaker Hershey trends and slower Honeywell Aerospace growth, alongside positive calls on Nvidia, Incyte and GE Vernova.

Analysis

The clearest signal is widening earnings-quality dispersion in semiconductors, not a sector-wide read-through. Continued AWS purchases of NVIDIA systems alongside Amazon’s in-house accelerators suggest custom silicon is currently additive to, rather than an immediate replacement for, merchant GPU demand. That supports NVIDIA near term; it does not establish durable order growth or margins. Intel faces the opposite test: PC weakness can offset data-center strength, while absent external foundry traction delays the strategic payoff and leaves execution risk in the multiple. Monolithic Power’s design-win ramps could benefit from data-center power intensity, but the cited 2026 growth rates concern Enterprise and Communications, not necessarily consolidated revenue; extrapolating them would be hazardous.

Over the next days, post-earnings positioning may dominate fundamentals. Over 1–3 months, verify AWS infrastructure spending, NVIDIA demand commentary, Intel foundry customer wins, and MPWR segment conversion into recognized revenue. Over 6–18 months, custom accelerators are a genuine substitution risk for GPUs, while power delivery remains a complementary content opportunity. The contrarian risk is treating analyst upgrades as fresh evidence: these calls largely interpret the same earnings and may already be reflected in prices. Apple, Amazon, and Roblox have event or guidance uncertainty; no valuation or consensus data here supports a high-conviction directional trade. Falsifiers include a material AWS spending slowdown, NVIDIA order/guidance deterioration, credible Intel foundry customer commitments, or MPWR ramp slippage.

Key Decisions for Investors

• Prefer a measured long-NVIDIA / underweight-Intel semiconductor expression rather than a broad chip-sector bet. Add only if post-earnings price action is orderly; thesis weakens on NVIDIA demand or AWS capex deterioration, and improves for Intel only with verifiable external foundry wins.
• Keep Monolithic Power on a pullback/watchlist, not as an automatic chase: confirm Enterprise and Communications design wins are ramping into reported revenue and check consolidated guidance before underwriting the cited segment growth rates.
• Treat Apple and Amazon as event-risk exposures into their next updates; monitor component-cost pressure, gross-margin commentary, AWS growth and spending rather than relying on broker target changes.
• Underweight Reddit only if traffic and monetization data confirm search-referral weakness; a single downgrade is not enough to establish a durable revenue impairment. Reassess on evidence of traffic stabilization or improved non-search acquisition.