5 big analyst AI moves: SpaceX gets a Street-high target, Samsung selloff overdone

Raymond James initiated SpaceX coverage with a Strong Buy and a $800 price target, projecting revenue rising from ~$38.5B and EBITDA of $17.7B today to $837B revenue and $696B EBITDA by 2031 on Starship’s cost reductions (transport cost -99%+) and payload scaling. Bank of America reiterated a Buy on Alibaba ($172/ADS) expecting AI-led cloud growth to accelerate to 45% YoY (from 38%) with cloud margins improving to 11% (from 9%) despite expected -7.7% YoY customer management revenue. In contrast, Bernstein downgraded Salesforce to Sector Weight from Outperform, citing weak feedback on Agentforce AI with IT budget deprioritization signals, while Stifel upgraded Shopify to Buy, targeting $150 and projecting 30%+ revenue growth in 2026. Overall, the analyst moves skew constructive on AI/cloud beneficiaries but cautious on AI monetization momentum for enterprise software.

Analysis

The cleanest read-through is not “AI is good,” but that AI spend is splitting the market into monetizers and narrators. Names with direct transaction or consumption hooks — SHOP and BABA — can show real operating leverage if usage converts into take-rate or cloud dollars, while workflow SaaS like CRM is vulnerable to a longer proving period because buyers now demand data readiness and measurable ROI before expanding budgets.

On semis, the tape should care more about the next two earnings prints than the recent prelim noise. If Samsung’s memory commentary and TSMC/ASML confirm tightening supply, that supports a second leg in DRAM/NAND and reinforces the AI capex cycle, which is constructive for AMD into its July event and for foundry/packaging exposure more broadly. The failure mode is simple: if cloud capex or memory pricing pauses, the market will quickly compress the AI premium in the whole supply chain.

The contrarian point is that the market may be too eager to punish CRM for weak agent feedback while underpricing how slowly enterprise AI workflows actually migrate. That is a setup for dead money rather than immediate collapse. By contrast, BABA’s AI/cloud angle could re-rate faster than consensus expects if margin leverage shows up even while core commerce remains soft; the key question is whether investors are still anchoring to old China macro narratives instead of the earnings inflection path.

Key Decisions for Investors

• Go long SHOP / short CRM for 1-3 months; thesis is share gains and monetization optionality versus a longer Agentforce digestion period. Target a 15-20% relative move; exit if CRM shows accelerating net-new deal momentum or SHOP growth decelerates materially below the current mid-20s/30s narrative.
• Buy BABA on weakness ahead of the next quarter as a cloud/AI margin repair story, not a China retail story. Risk/reward improves if cloud growth stays in the mid-40s and cloud margin expands; falsify on a cloud growth miss below ~40% or renewed QC loss escalation.
• Add AMD into the July 23 AI event only if semis are confirmed by TSMC/Samsung commentary. The setup is a favorable asymmetry into guidance; invalidate if TSMC/ASML imply capex digestion or if AMD’s data-center commentary comes without a revenue inflection.
• Watch/trim broad semi exposure if Samsung’s late-July memory commentary fails to validate the pricing rebound. The market is likely pricing in a stronger 2H26 than is currently visible; that makes the next two weeks a catalyst window, not a buy-and-forget setup.