Here are Thursday's biggest analyst calls: Nvidia, Tesla, Marvell, Broadcom, Dell, Dick's, Meta, Salesforce & more

Wall Street analysts issued a largely positive batch of calls, including upgrades for Quanta Services, LyondellBasell, Dominion, Dick's, Agilent, and Sterling Infrastructure, alongside reiterated positive views on Nvidia, Meta, Broadcom, Micron, Dell, Salesforce, Snowflake, and Tesla. Several notes highlighted AI/data center strength, Tesla robotaxi expansion, and continued demand tailwinds in semis and infrastructure. The most notable stock-specific move risk comes from price-target raises on Marvell to $240, Snowflake to $300, Micron to $1,150, Dell to $350, and Quanta to $800.

Analysis

The tape is still being driven by the same two-factor regime: AI capex winners and balance-sheet/asset-light “quality duration” stories. The biggest second-order implication is that the market is broadening the AI trade beyond semis into the infrastructure stack — power, networking, servers, and systems integration — which is usually when the cycle becomes self-reinforcing because backlog visibility starts to feed new capacity commitments. That said, this is also the stage where investors begin to overpay for anything adjacent to AI, so dispersion should rise sharply between true constraint beneficiaries and names merely labeling themselves as beneficiaries.

On the upside, the strongest setup is for companies where earnings power can inflect faster than consensus models: power/transmission, server OEMs tied to AI racks, and networking silicon. The risk is that the market is now capitalizing 2026-27 demand assumptions into today’s multiples, so any delay in hyperscaler spend, a softer order conversion rate, or margin pressure from customer concentration could produce 10-15% air pockets even if fundamentals remain sound. In other words, the winners are probably right, but the path is increasingly crowded and volatility should be used tactically rather than chased.

Outside AI, the cleaner theme is “asset quality plus operating leverage”: senior housing, infrastructure services, and select consumer names with event-driven demand. These are less about instant earnings beats and more about multiple expansion as investors pay up for visible growth and resilience. The contrarian miss is that several of these have already de-risked on the back of analyst upgrades, so the best risk/reward may now be in relative value rather than outright longs, especially where execution is strong but the market has only started to price it in.

Key Decisions for Investors

• Long NVDA / short CRM for 1-3 months: own the name with the clearest AI monetization and fade the software platform where AI remains more narrative than earnings; target 8-12% relative outperformance if AI capex keeps accelerating.
• Long DELL vs short a broad software basket for 2-6 months: DELL is one of the few ways to express server-rack leverage with near-term backlog conversion; risk is that order timing slips, but upside remains asymmetric if AI deployments broaden.
• Add PWR and STRL on pullbacks, with 6-12 month horizon: these are infrastructure leverage plays where backlog and pricing can translate into estimate upgrades; use a 10% trailing stop because multiples can compress quickly if rates spike.
• Buy a small NVDA/AVGO pair-long basket and hedge with a short in the highest-multiple AI-adjacent software names for the next earnings window: this captures the “picks and shovels” premium while reducing exposure to narrative-driven de-rating.
• If entering TSLA, prefer optionality over common stock: 6-12 month call spreads rather than outright long shares, since robotaxi expansion is a medium-dated catalyst and the stock can underperform sharply if the rollout pace disappoints.