Nvidia Earnings Blow Everyone Away

Nvidia reported 106% year-over-year revenue growth and guided for roughly 70% growth next year, versus Wall Street expectations of about 45%, with management indicating it can meet only around 70% of observed demand due to supply constraints. The outlook implies top-five hyperscaler capital spending could rise from roughly $800 billion this year to $1.3 trillion next year, supporting AI infrastructure suppliers such as memory, interconnect, inspection and data-center electrical-system providers while pressuring consumer-device manufacturers exposed to higher memory costs. CrowdStrike shares rose about 18% after net new ARR increased 51% year over year, backlog reached more than $10 billion (+49%), adjusted operating margin expanded 350bps, and management raised full-year guidance, though its roughly 37x sales valuation and stock-based compensation remain key risks.

Analysis

The investable read-through is not simply higher GPU volumes; it is a lengthening bottleneck chain. NVDA's incremental revenue requires advanced packaging at TSM, HBM supply, optical/DSP interconnect, rack integration and power delivery, creating greater operating leverage for CRDO, ONTO, IESC and, outside the supplied list, MU. NVDA already discounts an exceptional execution path, so second-derivative suppliers with smaller revenue bases offer better upside if orders convert, while TSM's upside is constrained by foundry-capacity allocation and geopolitical discount.

The principal near-term risk is that hyperscaler commitments represent multi-quarter capacity reservations rather than durable utilization demand. Over the next 1-3 months, watch cloud capex guidance, lead times, HBM pricing and NVDA gross-margin progression: a deceleration in any two would signal that supply is catching demand and pressure the high-beta infrastructure complex. Over 6-18 months, local inference is additive to total endpoint security demand but subtractive from the assumption that all AI value accrues to centralized compute; AAPL, DELL and enterprise edge vendors gain strategic relevance, while commodity PC assemblers remain exposed to memory-cost pass-through.

CRWD's reported momentum should not be extrapolated one-for-one into economic ARR because flexible consumption contracts can pull spending forward and defer churn visibility. Its valuation leaves little tolerance for a net-retention or GAAP-margin miss, particularly if stock-based compensation remains materially dilutive. Prefer ZS or RBRK for cybersecurity exposure: both retain AI-security optionality with less dependence on sustaining a premium multiple, although RBRK carries a shorter public operating history and higher execution risk.

Key Decisions for Investors

• Initiate a 3-6 month long CRDO / short NVDA beta-adjusted pair on a pullback: CRDO has greater revenue sensitivity to scale-out networking, while the NVDA short hedges broad AI-capex disappointment. Target 20-30% relative upside; exit if CRDO guide-downs its next-quarter revenue or hyperscaler networking orders soften.
• Accumulate ONTO and IESC in tranches over 1-3 months rather than chase post-earnings infrastructure momentum. The thesis requires sustained advanced-memory/packaging tool demand and data-center construction backlog growth; stop adding if book-to-bill falls below 1x or margins contract despite revenue growth.
• Rotate a portion of CRWD exposure into long ZS and/or RBRK for the next 6-12 months. This preserves the endpoint, zero-trust and data-resilience security theme while reducing multiple-compression risk; invalidate if CRWD sustains materially faster net-new ARR growth without incremental dilution or retention deterioration.
• Maintain a tactical short or underweight HP and GRMN versus DELL over the next two earnings cycles. Memory inflation is most damaging where pricing power is limited and AI-server mix is absent; cover if consumer-device gross margins hold flat or improve despite higher component costs.
• Set an NVDA risk alert around gross-margin guidance and TSM advanced-packaging capacity commentary. A meaningful margin reset or evidence of capacity normalization would be the earliest signal to reduce CRDO, ONTO and IESC, even if headline AI demand remains strong.