SMH Holdings: Complete Guide to the VanEck Semiconductor ETF (2025)
The VanEck Semiconductor ETF (SMH) is one of the largest and most-traded pure-play semiconductor funds in the United States, with over $47 billion in assets under management. Whether you're researching SMH's top holdings, evaluating NVIDIA's outsized weight, comparing SMH to SOXL or XSD, or looking for context on related tickers like MSOX and SOXL premarket moves — this guide covers it all.
What Is SMH? — VanEck Semiconductor ETF Overview
The VanEck Semiconductor ETF (ticker: SMH) is an exchange-traded fund that seeks to replicate the performance of the MVIS US Listed Semiconductor 25 Index. Managed by VanEck, the ETF was launched in December 2011 and has become the go-to vehicle for investors seeking pure-play exposure to the global semiconductor industry — including chip designers, manufacturers, and equipment suppliers.
Unlike broader technology ETFs such as QQQ or VGT, SMH is sector-concentrated: every holding is a semiconductor or semiconductor-adjacent company. This concentration amplifies both the upside during chip boom cycles and the downside during downturns, making SMH suitable primarily for investors with higher risk tolerance and a long-term thesis on AI, cloud computing, and digital infrastructure growth.
SMH Key Facts at a Glance
| Metric | Value |
|---|---|
| Full Name | VanEck Semiconductor ETF |
| Ticker Symbol | SMH |
| Issuer | VanEck |
| Inception Date | December 20, 2011 |
| Underlying Index | MVIS US Listed Semiconductor 25 Index |
| Number of Holdings | 25 |
| Expense Ratio | 0.35% |
| Assets Under Management (AUM) | ~$47 Billion (Feb 2026) |
| Average Daily Volume | High liquidity (top 1% of ETFs) |
| Distribution Frequency | Annual |
| Beta vs. S&P 500 | ~1.54–1.90 (high volatility) |
| Exchange | NYSE Arca |
Data sourced from VanEck, Morningstar, and StockAnalysis. Holdings and weights change at each quarterly rebalance. Verify current data before trading.
SMH Top Holdings & Weights
SMH tracks the 25 largest U.S.-listed semiconductor companies, weighted primarily by market capitalization — with a single-stock cap of 20% applied at each rebalance. This cap is critical context for understanding how NVIDIA's weight is managed (see the next section).
The table below reflects approximate holdings and weights as of early 2026. Weights fluctuate daily with price movements and are formally reset at each quarterly rebalance.
| Rank | Ticker | Company Name | Approx. Weight (%) | Sub-Sector |
|---|---|---|---|---|
| 1 | NVDA | NVIDIA Corporation | ~18.7% | Fabless / AI GPUs |
| 2 | TSM | Taiwan Semiconductor Mfg. Co. (ADR) | ~10.8% | Foundry / Contract Mfg. |
| 3 | AVGO | Broadcom Inc. | ~8.4% | Fabless / Networking |
| 4 | ASML | ASML Holding N.V. (ADR) | ~5.6% | Semiconductor Equipment |
| 5 | QCOM | QUALCOMM Incorporated | ~4.3% | Fabless / Mobile Chips |
| 6 | AMD | Advanced Micro Devices, Inc. | ~4.1% | Fabless / CPU / GPU |
| 7 | MU | Micron Technology, Inc. | ~3.8% | Memory (DRAM / NAND) |
| 8 | AMAT | Applied Materials, Inc. | ~3.6% | Semiconductor Equipment |
| 9 | LRCX | Lam Research Corporation | ~3.4% | Semiconductor Equipment |
| 10 | KLAC | KLA Corporation | ~3.1% | Process Control Equipment |
| 11 | TXN | Texas Instruments Inc. | ~2.9% | Analog / Embedded |
| 12 | ADI | Analog Devices, Inc. | ~2.7% | Analog / Mixed-Signal |
| 13 | MRVL | Marvell Technology, Inc. | ~2.5% | Fabless / Data Infrastructure |
| 14 | ON | ON Semiconductor Corp. | ~2.0% | Power / Automotive Chips |
| 15 | MCHP | Microchip Technology Inc. | ~1.8% | Microcontrollers |
| Remaining 10 positions constitute ~22% of the fund. Weights are approximate and change daily. | ||||
Source: VanEck, StockAnalysis.com. Holdings rounded to one decimal. Always check VanEck.com for the most current daily allocation.
Understanding SMH's Holdings Structure
Several factors stand out in SMH's portfolio construction:
- Top 3 concentration: NVDA, TSM, and AVGO together represent roughly 38% of the fund — meaning a major move in any of these three stocks significantly impacts SMH's daily performance.
- Equipment vs. Fabless split: About 25–30% of SMH is allocated to semiconductor equipment companies (ASML, AMAT, LRCX, KLAC), providing indirect exposure to the entire chip supply chain rather than just chip designers.
- Global geographic spread: Despite tracking U.S.-listed companies, SMH holds significant exposure to Taiwan (TSM) and the Netherlands (ASML), introducing geopolitical risk that pure domestic ETFs like XSD do not carry.
- No memory specialization: While Micron (MU) is included, SMH is not optimized for memory-centric investors — it skews heavily toward logic semiconductors and AI-related chips.
For investors who want broader exposure beyond the 25 largest names, the complete list of U.S.-listed semiconductor companies on InvestSnips offers a comprehensive directory of the full chip sector.
NVIDIA Concentration Risk & the 20% Cap Rule Explained
The single most important structural feature of SMH that investors frequently misunderstand is the 20% single-stock cap. Here's how it works in practice:
The MVIS US Listed Semiconductor 25 Index, which SMH tracks, applies a 20% weighting cap on any single security at each quarterly rebalance. NVIDIA's natural free-float market-cap weight within the index would routinely exceed this cap — sometimes by a wide margin during GPU bull markets. When this happens, VanEck must effectively sell some NVIDIA exposure at each rebalance to bring it back down to 20%, distributing that weight across other holdings.
Why This Matters for Your Investment
- Upside dampening: If NVIDIA rallies sharply between rebalances, SMH captures that gain intraday, but sells down NVIDIA at the next rebalance — potentially missing further appreciation at the top position.
- Built-in diversification benefit: The cap forces the fund to maintain meaningful allocations to names like AMD, Broadcom, TSMC, and equipment makers — preventing full NVIDIA dependency.
- Effective NVIDIA weight: In practice, NVIDIA's day-to-day intraday weight can drift above 20% before rebalance. Investors should treat approximately 18–22% as the realistic NVIDIA range within SMH at any given time.
Curious how SMH compares to AI-focused ETFs that hold NVIDIA at even higher weights? See InvestSnips' AI stock list for a curated breakdown of AI-heavy market vehicles.
SMH Performance History
SMH has delivered exceptional long-term returns driven by the global expansion of AI infrastructure, data centers, and advanced chip demand. Below is a summary of SMH's annual and cumulative performance across key timeframes.
| Period | SMH Total Return | S&P 500 Total Return | Context |
|---|---|---|---|
| 2023 | +74.7% | +26.3% | AI boom, NVIDIA multiple expansion |
| 2024 | +44.0% | +25.0% | Continued GPU demand, data center buildout |
| 2025 (est.) | +47.1% | ~+20% | Broadening AI chip demand |
| 3-Year (Ann.) | ~+55% | ~+24% | Semiconductor outperformance cycle |
| Since Inception (Ann.) | ~+27.7% | ~+14% | Dec 2011 – Feb 2026 |
Returns are approximate, sourced from VanEck, StockAnalysis, and Morningstar. Past performance does not guarantee future results. Annual returns are price return unless noted.
Despite its impressive track record, SMH has also experienced sharp drawdowns when semiconductor cycles turn: in 2022 it fell over 35% amid inventory corrections and rate hikes. This cyclicality is one of the fund's most significant risk factors and is discussed in depth below.
SMH vs. SOXL vs. XSD vs. MSOX — Full Semiconductor ETF Comparison
The supporting keywords for this topic — MSOX, SOXL, XSD stock, ETF with NVIDIA — reflect genuine investor research. Below is a comprehensive head-to-head comparison of the four most-searched semiconductor ETFs, with key differentiators called out for each.
| Attribute | SMH | SOXL | XSD | MSOX |
|---|---|---|---|---|
| Full Name | VanEck Semiconductor ETF | Direxion Daily Semiconductor Bull 3x | SPDR S&P Semiconductor ETF | AdvisorShares MSOS Daily Leveraged ETF |
| Leverage | 1x (No leverage) | 3x Daily | 1x (No leverage) | 2x Daily |
| Primary Exposure | 25 largest global semis (US-listed) | ICE Semiconductor Index (daily 3x) | ~46 equal-weighted US semis | US Cannabis sector (NOT semis) |
| Expense Ratio | 0.35% | 0.75% | 0.35% | ~0.98% |
| NVIDIA Weight | ~18–20% (capped) | Tracks index, amplified 3x daily | ~2–3% (equal weight) | Zero (cannabis ETF) |
| AUM | ~$47 Billion | ~$8–10 Billion | ~$1–2 Billion | Small (niche fund) |
| Holding Period Suitability | Long-term / core position | Short-term / traders only | Long-term / core position | Short-term / traders only |
| Geographic Exposure | US + Taiwan + Europe | Same as index (leveraged) | US-only companies | US cannabis (no semis) |
| Rebalancing | Quarterly | Daily (leverage reset) | Quarterly | Daily (leverage reset) |
| Best For | Broad semi sector conviction | Tactical semi sector bets | Equal-weight diversification | Cannabis sector only |
⚠ Important: MSOX (AdvisorShares MSOS Daily Leveraged ETF) is a 2x leveraged cannabis ETF — it has NO semiconductor exposure. It is sometimes mistakenly compared to semi ETFs due to ticker similarity. Verify the underlying before investing.
SMH vs. XSD: Market-Cap Weight vs. Equal Weight
This is one of the most underappreciated distinctions in semiconductor ETF investing. SMH is market-cap weighted — NVIDIA, TSMC, and Broadcom dominate. XSD is equal weighted — no single stock exceeds roughly 3–4% of the fund, giving Micron, ON Semiconductor, and smaller-cap names the same representation as giants.
- XSD advantages: Less NVIDIA concentration risk; broader exposure to mid-cap and small-cap semiconductor innovators; historically outperforms during sector rotations away from mega-caps.
- SMH advantages: Higher liquidity, larger AUM, stronger correlation to AI supercycle winners (NVDA, TSMC, AVGO); better suited if mega-cap semis lead the market.
- Verdict: Risk-tolerant investors bullish on AI chip leaders often prefer SMH; diversification-focused investors concerned about NVIDIA overconcentration may prefer XSD as a complement or alternative.
Browse InvestSnips' full list of U.S. semiconductor ETFs for a complete directory of all chip-sector funds available to U.S. investors.
SOXL Premarket & After‑Hours: What Investors Need to Know
Search volumes for "SOXL premarket" and "SOXL after hours" spike dramatically whenever major semiconductor news breaks — NVIDIA earnings, geopolitical chip restrictions, or macro Fed announcements. Understanding how leveraged ETFs behave during extended hours is essential risk management.
Why SOXL Moves So Aggressively Outside Regular Hours
SOXL seeks to deliver 3x the daily performance of the ICE Semiconductor Index. During premarket (4:00–9:30 AM ET) and after-hours (4:00–8:00 PM ET) sessions:
- Lower liquidity: Bid-ask spreads widen significantly. A 3x leveraged ETF trading with thin volume can move far beyond what the underlying index warrants.
- No real-time NAV reset: The fund's daily leverage reset happens at market close. Extended-hours prices reflect sentiment and supply/demand — not the fund's true daily leverage calculation.
- News amplification: A 5% move in NVIDIA during an after-hours earnings report can translate to a 12–18% swing in SOXL's after-hours price quote — though this is not equivalent to the fund's actual structured exposure.
- Gap-up/gap-down risk: If SOXL closes at $62 after regular hours but one of its major holdings announces a major miss overnight, the fund may open sharply lower the next day — compounding losses before the average investor can react.
How SMH Behaves vs. SOXL in Extended Hours
SMH, as a non-leveraged ETF with ~$47B in AUM and one of the highest trading volumes among sector ETFs, tends to trade with tighter bid-ask spreads and more accurate pricing even during extended hours. This makes SMH preferable for investors who need to act on overnight semiconductor news without accepting excessive slippage.
For investors wanting to understand the full semiconductor sector landscape, including related companies and their individual stock profiles, the InvestSnips complete list of semiconductor companies on U.S. exchanges is a valuable reference resource.
Risks & Downsides of Investing in SMH
Despite its strong long-term track record, SMH carries several distinct risk factors that investors must evaluate carefully before allocating capital:
1. Sector Concentration Risk
SMH holds exactly 25 semiconductors and semi-equipment stocks. There is no diversification into healthcare, consumer staples, financials, or other defensive sectors. When the semiconductor cycle turns — as it did sharply in 2022 — there is no buffer. SMH fell over 35% in 2022 due to inventory corrections, rate hikes, and slowing PC demand.
2. NVIDIA Single-Stock Risk
Even with the 20% cap, a severe decline in NVIDIA stock — whether from regulatory sanctions, chip export controls, or a major earnings miss — would directly hit approximately 18–20% of SMH's portfolio value in a single position. This level of dependency on one company is atypical even among sector ETFs.
3. Geopolitical / Taiwan Risk
Taiwan Semiconductor (TSM) represents approximately 10–11% of SMH and manufactures the world's most advanced chips, including NVIDIA's A100 and H100 GPUs. Any geopolitical escalation between China and Taiwan — or U.S. export control tightening — could sharply impact TSMC's operating prospects and, by extension, a significant portion of SMH's portfolio value.
4. Cyclicality & Inventory Correction Risk
The semiconductor industry is inherently cyclical. Boom periods of chip scarcity are followed by bust periods of inventory oversupply. SMH investors must be prepared for multi-year drawdown periods (the 2022 cycle saw >35% peak-to-trough losses) even during otherwise resilient broader markets.
5. Valuation Risk
As of 2025–2026, major SMH constituents — particularly NVIDIA and Broadcom — trade at elevated price-to-earnings multiples that reflect optimistic AI infrastructure spending projections. If those projections disappoint, valuation compression could lead to outsized SMH losses even without a fundamental deterioration in chip demand.
6. Currency and ADR Risk
SMH holds ADRs for TSMC (Taiwan) and ASML (Netherlands). Changes in USD/TWD or USD/EUR exchange rates can affect the value of these positions beyond the underlying company fundamentals.
Who Should (and Shouldn't) Invest in SMH?
✅ SMH May Be Suitable For:
- Long-term growth investors who believe the AI infrastructure buildout represents a multi-decade semiconductor demand cycle
- Thematic ETF investors seeking a single-fund way to access chip leaders — NVDA, TSMC, AVGO, AMD — without picking individual stocks
- Satellite portfolio allocation as a growth-oriented tilt within a diversified core-satellite portfolio (e.g., 5–15% of total portfolio)
- Sector-rotation traders who want liquid, low-cost access to the semiconductor industry during cyclical upswings
❌ SMH May Not Be Suitable For:
- Conservative or income-focused investors — SMH's dividend yield is minimal; it is a growth play, not an income vehicle
- Investors seeking broad diversification — its 25-stock, single-sector concentration is inappropriate as a core holding for conservative portfolios
- Short-term traders seeking leverage — SOXL (3x) or other leveraged instruments are more appropriate for tactical, high-conviction, short-duration bets
- Investors overconcentrated in NVIDIA — if you already hold large positions in NVDA individually, adding SMH further concentrates your NVIDIA exposure
Looking at how semiconductor stocks fit within the broader S&P 500 technology allocation? See InvestSnips' S&P 500 technology stocks guide to understand how the chip sector sits within the broader tech universe.
How to Evaluate a Semiconductor ETF: 6-Point Framework
Whether you're deciding between SMH, SOXL, XSD, or any other chip-sector ETF, use this six-point evaluation framework to make an informed comparison:
| Evaluation Criterion | What to Assess | SMH Profile |
|---|---|---|
| 1. Expense Ratio | Lower is better for long-term holders. Under 0.5% is strong for sector ETFs. | 0.35% — Competitive ✅ |
| 2. Liquidity & Volume | Higher daily volume = tighter spreads = lower trading cost. | Top-tier liquidity; one of the most-traded sector ETFs ✅ |
| 3. Concentration Risk | Top-10 holdings as % of total. Over 60% = elevated single-stock risk. | Top 3 = ~38%. High concentration — monitor ⚠ |
| 4. Geographic Exposure | International holdings add geopolitical and currency risk. | TSMC (Taiwan), ASML (Netherlands) — meaningful geo-risk ⚠ |
| 5. Index Methodology | Market-cap vs. equal-weight changes risk profile dramatically. | Capped market-cap weight (20% max per stock) — balanced ✅ |
| 6. Leverage & Structure | Is it 1x, 2x, or 3x? Leveraged ETFs are inappropriate for long holds. | 1x — no leverage; appropriate for long-term investors ✅ |
To explore more ETFs beyond semiconductors, InvestSnips maintains a full list of U.S. technology-focused ETFs spanning AI, cloud, cybersecurity, and more.
Summary & Key Takeaways
- 📌 SMH is the dominant semiconductor ETF with ~$47B AUM, tracking the 25 largest U.S.-listed semiconductor companies via the MVIS US Listed Semiconductor 25 Index.
- 📌 NVIDIA is capped at ~20% at each quarterly rebalance — understanding this cap mechanism is critical for modeling SMH's NVIDIA sensitivity.
- 📌 Top 3 holdings (NVDA + TSMC + AVGO) represent approximately 38% of the fund, creating meaningful concentration risk in mega-cap chip names.
- 📌 SMH expense ratio of 0.35% matches XSD and is lower than leveraged alternatives like SOXL (0.75%).
- 📌 MSOX is a cannabis ETF, not a semiconductor ETF — despite frequent confusion with MSOS and similar tickers.
- 📌 SOXL premarket and after-hours trading carry elevated risk due to low liquidity, wide spreads, and leverage compounding — avoid extended-hour trades in leveraged ETFs without a specific strategy.
- 📌 SMH is best held long-term within a diversified portfolio as a thematic sector play — not as a core sole-holding.
- 📌 XSD offers equal-weight alternative for investors concerned about mega-cap concentration, at the same 0.35% expense ratio.
Frequently Asked Questions About SMH Holdings
SMH rebalances quarterly, which is when the MVIS US Listed Semiconductor 25 Index is formally reconstituted. During each rebalance, the 20% single-stock cap is applied, which typically results in VanEck trimming NVIDIA's weight if it has grown above the cap. Between rebalances, individual stock weights drift daily with market price movements. Investors should check VanEck's website for the current daily weight disclosures.
SMH can be held in a tax-advantaged retirement account like an IRA or 401(k), and its long-term growth profile has historically rewarded patient investors. However, its high concentration in a single cyclical sector means it can experience drawdowns of 30–40% during chip downturns, which may be uncomfortable for investors approaching retirement. A common approach is using SMH as a satellite position (5–15% of a portfolio) within a broader diversified retirement allocation, not as a primary holding.
SMH (VanEck) tracks the MVIS US Listed Semiconductor 25 Index, holding 25 companies with a 20% single-stock cap. SOXX (iShares) tracks the ICE Semiconductor Index and holds 30 companies, applying a different weighting methodology. Both are market-cap weighted and have similar expense ratios, but differ in geographic exposure, number of holdings, and top-weight construction. SMH has significantly higher AUM (~$47B vs. ~$23B for SOXX), providing greater liquidity for institutional and active traders.
SOXL is designed explicitly for short-term, tactical trading — not long-term buy-and-hold investing. Its 3x daily leverage reset creates a phenomenon called "volatility decay" (also known as beta slippage), where holding a leveraged ETF through volatile periods results in returns that lag 3x the underlying index over multi-month or multi-year periods. In a volatile semiconductor environment, a long-term SOXL holder can significantly underperform even a 3x multiple of SMH's unleveraged returns. SMH is the appropriate vehicle for long-term semiconductor exposure.
MSOX — the AdvisorShares MSOS Daily Leveraged ETF — is a 2x leveraged cannabis ETF with zero exposure to semiconductors. The ticker similarity to MSOS (AdvisorShares Pure US Cannabis ETF) and the fact that it appears in searches alongside semiconductor-themed ETFs creates confusion. MSOX is sometimes grouped with semiconductor searches by algorithmic search clustering, but investors must verify the underlying index before placing any trade. MSOX and SMH are completely unrelated investment vehicles.
XSD (SPDR S&P Semiconductor ETF) uses an equal-weighting methodology, meaning all ~46 holdings carry roughly the same weight (~2–3%) at rebalance. This significantly reduces NVIDIA concentration risk and gives greater weight to mid-cap and small-cap semiconductor names. XSD may outperform SMH during periods when smaller chip companies lead the sector, and may underperform when mega-cap names like NVDA and TSMC are the primary drivers. Both carry a 0.35% expense ratio, making cost-of-carry identical — the choice comes down to cap-weight conviction versus equal-weight diversification preference.
SMH does pay a small annual dividend, but its yield is minimal — typically well under 1%. The fund's constituents (led by NVIDIA, TSMC, Broadcom) are predominantly growth-oriented companies that reinvest earnings rather than returning large amounts as dividends. Semiconductor ETFs like SMH are growth vehicles, not income vehicles. Investors seeking significant dividend income should look to dividend-focused ETFs or individual high-yield stocks rather than SMH. For dividend stock ideas, see InvestSnips' top dividend stocks guide.
Several macro and sector-specific factors could drive SMH underperformance: (1) A slowdown in AI data center capital expenditure from hyperscalers like Microsoft, Google, or Amazon; (2) U.S. export control tightening on advanced chips to China, directly impacting NVIDIA and TSMC revenue; (3) A global semiconductor inventory correction, similar to 2022, as order backlogs normalize; (4) Geopolitical escalation in the Taiwan Strait, threatening TSMC's production capacity; (5) Rising interest rates, which compress growth stock valuations and hit high-multiple semiconductor names. These risks are not predictions, but they represent the primary scenarios that have historically caused sharp SMH drawdowns.