Here are the 2 big things we're watching in the stock market this week

Nvidia’s upcoming print is expected to anchor the AI complex, with LSEG calling for EPS of $2.09 and revenue of $92.04B, while the Fed’s Jackson Hole speech by Kevin Warsh (first as chair) and July PCE inflation (core PCE expected +0.18% m/m, +3.2% y/y) set the macro backdrop for rate expectations. On the trade front, U.S.–Canada talks collapsed and the U.S. imposed 50% tariffs on certain Canadian goods, with Canada’s retaliatory duties due to start Sept. 8—adding geopolitical risk. In company-specific earnings, Salesforce is forecast at $11.32B revenue and $3.27 adjusted EPS (with scrutiny on AI-era demand and fiscal 2027 Q3 guidance), while CrowdStrike is expected to report $1.44B sales and 29 cents adjusted EPS as cybersecurity demand for AI adoption remains key.

Analysis

This week is less about headline beats than about who can still compound under a tighter capital-cost regime. For NVDA, the key variable is whether pricing power can absorb input-cost inflation and still defend margins; if not, the market will quickly re-rate the broader AI supply chain because the financing burden is starting to leak from hyperscalers into vendors. That makes the largest upside asymmetry for the handful of names that can self-fund growth and the first downside for second-tier hardware suppliers with weaker leverage to price and mix.

Software is finally getting a chance to prove it is not just a valuation hostage to lower rates. The differentiator is whether AI is additive to seats, RPO and renewal rates, or merely a new bundle on top of slowing core demand; that is why cyber looks structurally better than horizontal SaaS. CRWD and PANW are the cleaner expressions because every new AI workflow expands the attack surface and raises security budgets, while CRM needs evidence that agentic products convert into net-new ARR rather than cannibalizing existing spend.

Jackson Hole is a duration event, not just a macro event. A hawkish reaction function would hit long-duration software and unprofitable tech first, while flatter curve beneficiaries like CME should outperform on volatility and rate uncertainty. The consensus may be overpricing the idea that AI hardware remains the only winner; the more durable 6-18 month trade is likely security plus the software vendors that can prove monetization, not just demos. The Canada tariff noise is probably second-order unless it bleeds into retail gross margins or border-sensitive industrial supply chains over the next 1-2 quarters.

Key Decisions for Investors

• Long NVDA only if guidance confirms margin resilience; if gross margin slips meaningfully below the mid-70s or forward revenue merely matches consensus, fade rallies above $210 over the next 1-3 trading sessions.
• Add CRWD on post-earnings strength for a 2-6 week hold; thesis breaks if guidance or next-quarter ARR growth fails to reaccelerate, in which case the stock likely retraces the entire event move.
• Use CRM as a show-me name into Dreamforce: buy only on a clean revenue/RPO beat with raised forward commentary; otherwise sell strength after earnings and rotate into CRWD/PANW.
• Long CME as a Jackson Hole/PCE hedge for the next 1-4 weeks; this works best if core PCE surprises hot or Warsh sounds more restrictive than the market expects.
• Avoid chasing the broad software basket until after Jackson Hole; if rates back up, the highest-beta SaaS names likely underperform even when fundamentals are fine.