Is the AI Trade Entering Its Second Wave? This ETF Could Be a Major Beneficiary.

The VanEck Semiconductor ETF (SMH) has returned more than 350% over the past three years and is positioned to benefit from the broadening AI-driven semiconductor boom. The fund is heavily top-weighted, with Nvidia at 16%, TSMC at 9%, Intel at 8%, and AMD at 7%, which should help if megacap chip leaders regain momentum. The article is constructive on SMH versus equal-weight alternatives, but it is largely an opinion piece rather than new market-moving information.

Analysis

The market is transitioning from a narrative trade to an earnings-discrimination trade, which should widen dispersion inside semis rather than lift everything uniformly. The biggest implication is that AI capex is no longer just a demand signal for designers; it is now filtering through the entire toolchain, with equipment vendors and foundry-adjacent names likely to see a longer runway in bookings even if unit growth in leading GPUs eventually moderates. That favors the picks-and-shovels cohort over late-cycle consumer-tech beta, because the next leg is about capacity expansion, advanced packaging, and node transitions rather than pure sentiment.

The key second-order effect is that SMH’s concentration makes it a cleaner expression of “leaders keep leading” but also more fragile if the market starts questioning marginal ROI on AI infrastructure. If hyperscaler spending pauses for even one quarter, the ETF’s top-heavy structure can underperform more balanced semis exposure as the market rotates to second-tier beneficiaries that are less directly tied to Nvidia-like momentum. In other words, the same concentration that helped SMH outperform can become a headwind if capex breadth broadens but megacap multiple expansion stalls.

The contrarian miss is that the market may be underpricing supply-chain bottlenecks in advanced packaging, lithography, and wafer fab tooling. Those constraints can keep pricing power elevated for 12-24 months even if end-demand growth normalizes, supporting margins for ASML, AMAT, and LRCX more persistently than consensus expects. The real risk is not near-term demand collapse; it is a second-half-2026 digestion phase where ordering slows before revenue catches up, compressing multiples across the group.

Key Decisions for Investors

• Go long ASML/AMAT/LRCX versus SMH on a 6-12 month horizon: better asymmetry if the market re-rates scarcity and tool demand, with lower downside if AI capex broadens but top-end growth slows.
• Use any post-earnings pullback in NVDA to own Jan-2027 calls or call spreads: upside remains tied to ecosystem leadership, but structure the trade to limit multiple-compression risk if growth decelerates over the next 2-3 quarters.
• Pair long TSM vs short INTC for a 3-6 month trade: the foundry/model shift should capture incremental AI demand while legacy manufacturing faces execution and mix headwinds; this is the cleaner expression of capacity monetization.
• Avoid chasing SMH outright after strong runs; if long, hedge with a small short in XSD or reduced beta exposure because equal-weight semis can outperform if leadership narrows and megacap multiples stall.
• Watch for commentary on hyperscaler capex and advanced packaging lead times at the next earnings cycle; if order growth inflects down, take profits on equipment longs and rotate into cash-rich designers.