The VanEck Semiconductor ETF, trading as SMH, is the most concentrated of the major chip funds and has delivered the strongest five‑year average annual return among the three ETFs compared.
Over the past five years SMH posted a 36% average annual return versus 31% for the iShares Semiconductor ETF, SOXX. SMH tracks the 25 stocks in the MVIS US Listed Semiconductor 25 Index and is market cap‑weighted with essentially no cap on individual positions; that approach has left the fund with a heavy mega‑cap tilt.
SMH’s top holding, Taiwan Semiconductor Manufacturing, accounts for 9.78% of the fund, and SMH’s combined positions in Nvidia and TSMC add up to roughly 25% of the portfolio. That concentration has amplified gains while making the fund more dependent on a handful of mega winners.
By contrast SOXX holds 30 names, tracks the ICE Semiconductor Index and cap‑weights its portfolio while limiting how much any one stock can account for. The limits give SOXX a slightly higher tilt to smaller companies than SMH and reduce single‑name exposure.
SOXQ, the Invesco PHLX Semiconductor ETF, also holds around 30 stocks and tracks the PHLX Semiconductor Sector Index; its portfolio is substantially the same as SOXX’s. SOXQ’s expense ratio is 0.19%, lower than SOXX’s 0.34%, and that cheaper fee has allowed SOXQ to modestly outperform SOXX over the past several years.
The timing of these differences matters now because capital spending by large tech companies has reshaped demand for chips. Microsoft, Amazon, Alphabet, Meta Platforms and Oracle committed nearly $700 billion in capital expenditures for 2026 — an 81% increase over the prior year — and the majority of that spending will help them meet semiconductor demand.
Semiconductor revenue itself reached $298.5 billion in the first quarter of 2026, up 25% from the fourth quarter of 2025, and IDC’s April 2026 forecast predicted the semiconductor market would exceed $1 trillion in revenue by the end of 2026. So far in 2026 semiconductors and the physical infrastructure of AI have dominated the ETF scene; the iShares Semiconductor ETF was up 89% year to date.
Those numbers explain why a concentrated, cap‑weighted fund like SMH has outperformed on average: when a small group of companies captures the bulk of sector gains, a fund that tilts into them rises faster. But the tradeoff is clear — concentration can cut both ways if leadership shifts or specific names pull back.
That tension matters to individual investors choosing among nearly identical name lists but different structures. If your goal is maximal exposure to the largest industry winners, SMH stock’s structure has delivered the best five‑year result. If you prefer lower fees and a cap that limits single‑stock risk, SOXQ or SOXX may better fit your case.
Eddie Ghabour warned investors to "prepare for choppy trading and look for buying opportunities," a reminder that even in a hot upcycle volatility can arrive suddenly and affect concentrated holdings hardest.
The unresolved question now is whether the AI and data‑center spending surge will continue to favor the most concentrated funds. The answer hinges on the durability of that near‑term capex and whether a handful of mega‑cap chip names keep leading revenue growth; investors deciding between concentration and cost should weigh that dependency against fees and allocation rules.
