European chip stocks extend Broadcom-driven selloff By Investing.com

Broadcom's disappointing AI-chip demand and unchanged fiscal 2027 AI revenue target of $100 billion triggered a broad selloff in semiconductors. The PHLX semiconductor index fell 10.3% on Friday, its worst one-day drop since March 2020, while Nvidia lost roughly 6% and more than $300 billion in market value; Micron dropped 13%, Marvell 17%, and AMD nearly 11%. European chip stocks extended the rout on Monday, with ASML, ASM International, BE Semiconductor, ams-OSRAM, and Siltronic down 1% to 3%.

Analysis

This is less a Broadcom-specific reset than a positioning unwind in the highest-duration part of semis. The market had been paying for an accelerating AI capex curve; a guidance plateau at the ecosystem anchor is enough to force de-grossing across the entire AI basket, especially names with elevated expectations and crowded ownership. The immediate second-order effect is that suppliers with the most leverage to near-term hyperscaler spend — custom ASICs, HBM, and advanced packaging — are likely to see multiple compression even if their own fundamentals have not changed.

The more important read-through is relative. If custom AI demand is merely shifting out in time rather than breaking, the market will eventually discriminate between platform winners with backlog visibility and “hope” beneficiaries priced on 2026–27 narratives. That should favor the most diversified compute exposure and penalize narrower adjacent names first, because they have less ability to absorb any one customer’s digestion cycle. In other words, the selloff can persist even if AI capex remains structurally intact, simply through lower confidence in the slope of incremental revisions.

Risk is two-sided but asymmetric over the next 2–6 weeks: momentum and systematic flows can keep pressure on semis longer than fundamentals justify, while a single strong order commentary from a hyperscaler or improved lead-time data could trigger a sharp bear-market rally. The key catalyst is not another earnings beat, but evidence that AI infrastructure spend is broadening beyond one vendor’s custom-chip pipeline. Until then, the burden of proof sits with bulls, and the path of least resistance is lower for the most consensus-long names.

The contrarian view is that the reaction may be overdone versus the actual signal content. Holding a long-term revenue target steady is not the same as cutting it, and the market may be extrapolating one datapoint into a full-cycle demand deceleration. If this is simply a digestion quarter, the best long after the shakeout may be the names with the strongest balance sheets and the most tangible share gains from weaker competitors rather than the highest-beta AI proxies.

Key Decisions for Investors

• Short NVDA / MRVL basket against a long ASML hedge only if the tape fails to stabilize for 3-5 sessions; thesis is continued multiple compression in the most crowded AI duration names, with ASML used as partial hedge because its end-market exposure is broader and less sentiment-sensitive.
• Buy 2-4 week put spreads on AMD or MRVL into any intraday bounce; these names have the highest sensitivity to flow-driven de-risking and offer better convexity than outright shorts if the market snaps back violently.
• Initiate a tactical long in ASML only after a further 5-7% drawdown or a clear reversal in semicap leadership; the risk/reward improves if investors begin rotating from AI beta into quality equipment exposure.
• Pairs trade: long AVGO less NVDA for 1-3 months if you expect the market to re-rate toward diversified platform cash flow versus single-theme AI exposure; the relative trade benefits if AI spending remains real but less explosive than priced.
• Watch for a catalyst-based short-covering setup in MU on any commentary confirming HBM demand resilience; if that data emerges, cover fast because memory names can rebound sharply once the market stops extrapolating a capex pause into a full demand collapse.