The article compares three semiconductor ETFs and argues that SOXQ is the best value, with a 0.19% expense ratio versus SOXX at 0.34% and broadly similar holdings. SMH is the most concentrated and has the strongest 5-year performance at 36% average annual return, while SOXX is more diversified but less cost-efficient. The piece is constructive on semiconductors overall, citing nearly $700 billion in 2026 capex from major tech firms that should support chip demand.
This reads less like a generic semis bull case and more like a dispersion trade inside the group. The incremental capital wave from hyperscalers should not lift all names equally: it should concentrate near-term earnings revisions in the fastest levered suppliers to advanced logic, HBM, and packaging, while leaving legacy/data-center adjacencies with lower beta to the capex cycle. In practice, that favors NVDA, TSM, MU, and to a lesser extent AMD and AVGO over broader “semi basket” exposure if the goal is to capture the first-order spend impulse.
The second-order effect is that ETF choice matters most when breadth narrows. If the AI buildout remains mega-cap led, a concentrated vehicle will outperform a more equalized basket because the revenue pool is becoming increasingly anchored to a handful of platform winners; if capex broadens into networking, analog, and industrial semis, the more diversified funds catch up. That means the market’s real decision point is not semis versus no semis, but whether the next leg is driven by a few AI accelerators or by a wider replacement cycle across the stack.
The contrarian risk is that the capex headline is already a known narrative and could be a “sell the announcement” event for the highest-multiple beneficiaries if utilization or deployment timing slips. Semis are also vulnerable to a 2H26 digestion phase: orders can look strong while customer inventories and take rates normalize, causing multiple compression even if revenue still grows. The key reversal catalyst would be any evidence that spend is shifting from buildout to optimization, or that one of the mega-cap buyers defers compute deployment by a quarter or two.
From a flow standpoint, the cleaner expression is not chasing the most crowded ETF, but owning the operational winners where estimate revisions can still outrun valuation. SMH is the higher-beta wrapper on the same theme, but if mega-caps de-rate, the cheaper basket should hold up better because concentration risk is lower and the market has already paid for the AI scarcity premium. The best risk/reward over the next 3-6 months is likely in names with direct exposure to advanced packaging and memory tightness, where supply constraints can keep pricing power elevated longer than the market expects.
AI-powered research, real-time alerts, and portfolio analytics for institutional investors.
Request DemoOverall Sentiment
mildly positive
Sentiment Score
0.15
Ticker Sentiment
