Micron, Marvell, and Broadcom sink, leading chip stocks lower

Technology and semiconductor stocks sold off sharply, with Broadcom down 14% after weaker-than-expected earnings and Micron, Marvell, AMD, Qualcomm and Intel also lower. Analysts said the pullback reflects elevated expectations after a strong AI-driven run, with concerns about falling chip prices and slower AI spending adding pressure. The Nasdaq fell 200 points in early trading, suggesting broader sector-wide risk-off sentiment.

Analysis

This is less about a one-day earnings miss than a positioning reset in the AI hardware complex. The first-order pressure is on names that have been priced for near-flawless execution, but the second-order effect is broader: capital is likely to rotate from “AI beneficiaries with stretched expectations” into the picks-and-shovels winners with clearer demand visibility, such as memory, networking, and foundry-capacity beneficiaries that can still underwrite earnings growth without multiple expansion. The fact that the move is led by a custom-chip supplier also matters for the ecosystem: hyperscalers have more leverage than the market had been pricing, which can compress supplier margins even if AI capex stays elevated.

The near-term risk is not an AI spending collapse; it is a digestion period where order growth decelerates faster than consensus models, causing estimate cuts and multiple compression to compound each other. That can persist for 4-8 weeks and create additional downside if sell-side revisions begin to lag spot pricing. The cleaner tell to watch is whether weakness stays confined to compute-exposed semis or spreads into broader tech: if software and internet megacaps hold up, this becomes a sector-specific de-rating; if they crack too, the market is repricing the entire AI trade as crowded and over-owned.

The contrarian view is that the pullback may be more of a leadership shakeout than a fundamental inflection. Many institutional portfolios are structurally underexposed to semis after chasing higher-beta leaders, so forced de-risking can overshoot fair value before long-only buyers return. If AI demand is merely slowing from hypergrowth to strong growth, the best risk/reward may be in names with the cleanest backlog and the most pricing power rather than the highest narrative beta.

Key Decisions for Investors

• Tactically short AVGO into the next 1-2 weeks of post-earnings revision risk; use a tight stop on any evidence of stabilizing customer order commentary. Risk/reward favors a 5-8% additional drawdown if estimate cuts follow, versus limited upside until sentiment resets.
• Pair long MU / short AVGO for 1-3 months. The market is likely to reward clearer supply-demand normalization and memory pricing leverage while penalizing custom-AI exposure where expectations are now the main vulnerability.
• Buy MRVL on weakness only after the next sell-side reset, not on the first bounce. The thesis is a cleaner rebound if the tape confirms this is a valuation air-pocket; upside can be 10-15% from depressed levels, but only after the de-rating runs its course.
• Use QCOM or INTC as relative shorts versus a broader tech basket for 2-4 weeks if risk-off persists. They have less direct AI optionality, so they can lag in a sector-wide unwind without needing company-specific bad news.
• For hedging, consider a short-dated NDAQ put spread as a proxy for broader tech risk-off. It offers cleaner convexity if the selloff migrates from semis into index-level de-risking.