Nvidia Delivers and the Nasdaq Jumps 1.3%

Stocks rose sharply as Nvidia and Salesforce blew past estimates, with NVDA surging 8.4% after posting $96.2B Q2 revenue (+106% YoY) and issuing strong Q3 guidance. The rally also lifted the Nasdaq +1.31% vs the S&P 500 +0.66% and Dow +0.33%, though participation was thin as most other names drifted lower. Sentiment remains vulnerable ahead of Friday’s Jackson Hole remarks after Cleveland Fed President Beth Hammack urged raising rates, with oil up 1% on Strait of Hormuz concerns and reports of new semiconductor/compute tariffs.

Analysis

This is a classic earnings-driven factor squeeze: the market is paying up for scarce growth with visible capex conversion, while everything tied to duration or domestic cyclicality is losing relative attention. The second-order winner is not just NVDA, but the AI infrastructure complex — advanced packaging, HBM memory, networking, and hyperscaler capex beneficiaries — because strong demand validation reduces the odds of an 1H27 capex pause. The catch is that if rates reprice higher, the market will start discounting that same AI spend at a harsher hurdle rate, which hits the rest of software first and then the semis multiple.

CRM’s move looks more like a sentiment rebound than a clean fundamental inflection: agentic AI can improve retention and pricing power, but it also raises the risk that model vendors capture the economic rent while application-layer software pays higher inference costs. That makes the near-term trade more about multiple expansion than near-term EPS revision, and it is fragile if Treasury yields back up or if management teams fail to show measurable AI monetization over the next 1-2 quarters. By contrast, NVDA has near-term operating leverage, but the reported memory-cost squeeze is the first credible reminder that even best-in-class suppliers can see gross margin giveback before pricing power reasserts itself.

The contrarian setup is breadth: a rally led by a handful of mega-caps is usually easier to fade at the index level than at the single-name level. If Jackson Hole turns hawkish, the first unwind is likely in high-multiple software and small caps, not in the earnings winners, so relative value matters more than outright beta. Over 6-18 months, the market’s real question is whether AI spend turns into enough productivity to offset rate pressure; until then, this looks like a narrow leadership market, not a healthy risk-on regime.

Key Decisions for Investors

• Long NVDA / short IWM for the next 1-3 months: best expression of narrow leadership and rate sensitivity, with the thesis failing if Friday’s Fed messaging pushes yields down and small caps catch up.
• Buy QQQ vs. SPY on dips only, but keep sizing modest: the index is being pulled by a few winners, so upside is concentrated while breadth risk remains; reduce if NVDA fails to hold post-earnings gains within 2-3 sessions.
• If holding CRM, pair it against a higher-duration software basket (e.g., XLK peers with weaker FCF conversion) rather than owning it outright: the market is rewarding AI narrative now, but the first real test is next quarter’s evidence of monetization.
• Set an alert on NVDA gross margin commentary and HBM pricing: if margins keep compressing beyond the next 1-2 quarters, the stock’s multiple can de-rate even if revenue stays strong.
• Avoid chasing semis broadly ahead of any tariff headline risk; wait for confirmation that tariffs are targeted/exempted before adding exposure to SMH/SOXX.