Broadcom's Stock Sinks Despite Solid Earnings. Other Chip Stocks Are Sliding Too.

Broadcom fell nearly 15% despite beating fiscal Q2 estimates, reporting adjusted EPS of $2.44 and revenue of $22.19 billion, up 49% year over year. The company forecast fiscal Q3 revenue of $29.4 billion, but kept its long-term outlook unchanged, disappointing investors who wanted a higher AI chip sales target beyond $100 billion for 2027. The weakness spilled into other chip stocks, including AMD, Nvidia, Intel, Arm, and Marvell.

Analysis

The tape is telling us this is no longer a fundamentals-only trade; it’s a positioning and narrative reset around AI capex durability. Broadcom was acting as a proxy for “AI monetization with margins,” so a guide that did not meaningfully extend the terminal growth story is enough to trigger de-risking across the whole semi complex, especially names with crowded AI ownership and high beta to the same factor basket. That makes the first-order loser AVGO, but the bigger second-order effect is multiple compression for the entire custom-silicon ecosystem as investors reprice how much of the AI buildout can be captured without sacrificing profitability.

The key mechanical risk is that custom silicon is a volume story before it is a margin story. If the market decides the AI infrastructure cycle is moving from scarcity pricing to procurement discipline, the winners rotate from “design wins” to “balance sheet and installed base,” which is a relative advantage for NVDA and the highest-quality platform names over smaller custom-chip exposure. MRVL and AMD are more vulnerable here because they sit closer to the “next increment of AI share gain” narrative; once the market questions whether those gains are already in the price, their upside asymmetry deteriorates quickly.

The move looks overdone tactically, but not necessarily strategically. In the next 1-3 sessions, forced selling and systematic de-grossing can keep pressure on the group irrespective of fundamentals; over 1-3 months, a stabilization in AI capex commentary should allow AVGO to reclaim some of the drawdown if margin fears prove too conservative. The bigger contrarian miss is that investors may be conflating “no higher guide” with “cycle peak,” when in reality the base case may simply be a normalization from extreme expectations rather than an imminent demand air pocket.

Key Decisions for Investors

• Buy AVGO on weakness only if it holds the post-earnings gap for 1-3 sessions; target a 10-15% rebound over 4-8 weeks, with a tight stop if the stock loses the drawdown low on heavy volume.
• Pair trade: long NVDA / short MRVL for 1-2 months. Thesis: platform/control-point exposure should outperform custom-silicon names if the market re-rates AI monetization toward quality and away from incremental share gains.
• Sell a 30-45 DTE AVGO put spread instead of outright long stock if you want to monetize volatility while keeping defined risk; best structure if implied vol remains elevated after the selloff.
• Reduce AMD and ARM tactical exposure until the market sees the next AI spending checkpoint; these names are most sensitive to a second derivative slowdown in AI enthusiasm and can underperform 5-10% on sentiment alone.
• If looking for a hedge, short SMH against long-book AI beta for the next 2-4 weeks; the goal is to isolate whether this is sector-wide de-rating versus company-specific disappointment.