Is This ETF the Smartest Investment You Can Make Today?

The VanEck Semiconductor ETF (SMH), which tracks the MVIS US Listed Semiconductor 25 index, has outperformed in 2025—up 38% year-to-date versus 17% for the Nasdaq—and concentrates nearly half its weight in five names: Nvidia (18.5%, 34.99M shares), Taiwan Semiconductor (9.5%), Broadcom (8.1%), Micron (6.8%) and AMD (6.6%). These holdings supply the GPUs, CPUs, DRAM, accelerators and foundry capacity that underpin AI data centers (Broadcom recently agreed to supply custom accelerators to OpenAI), making SMH a vehicle for diversified exposure to AI infrastructure while capping single-stock risk at 20% weightings. With a 0.35% expense ratio ($35 per $10,000) and a backdrop of forecasted AI market growth from $279bn in 2024 to $3.5tn by 2033 (CAGR ~31.5%), the ETF is pitched as a cost-effective, concentrated way to capture secular AI-driven semiconductor upside, though periodic corrections should be expected.

Analysis

The VanEck Semiconductor ETF (SMH) has outperformed in 2025, rising 38% year-to-date versus a 17% gain for the Nasdaq Composite, and it tracks the MVIS US Listed Semiconductor 25 index. The fund holds nearly 35.0 million shares of Nvidia (18.5% weighting) and sizable positions in Taiwan Semiconductor (9.5%), Broadcom (8.1%), Micron Technology (6.8%) and AMD (6.6%); those five names account for nearly half the ETF’s weight.

Holdings align directly with AI infrastructure needs: Nvidia supplies data-center GPUs, AMD provides CPUs for servers, Broadcom designs AI accelerators, Micron supplies DRAM for AI workloads, and TSMC provides foundry capacity. Broadcom’s announced collaboration with OpenAI for 10 gigawatts of custom accelerators is a concrete demand signal for accelerator supply.

The ETF charges a 0.35% expense ratio (approximately $35 per $10,000) and caps any single stock at 20%, leaving Nvidia close to the cap—concentration and idiosyncratic stock moves remain primary risks. Sentiment is moderately positive (score ~0.6) and the article cites a Grand View Research forecast of AI growth from $279bn in 2024 to $3.5tn by 2033 (CAGR ~31.5%), supporting a secular case but also the likelihood of periodic corrections.

Key Decisions for Investors

• Consider adding SMH for diversified, long-duration exposure to AI infrastructure given the ETF’s 38% YTD outperformance and the cited secular AI growth forecast
• Size positions conservatively and cap exposure because five stocks constitute nearly half the fund and Nvidia already represents 18.5%, approaching the 20% single-stock limit
• Monitor near-term catalysts such as Broadcom’s OpenAI accelerator deal, shifts in Nvidia weighting relative to the cap, and Micron demand trends as potential re-rating triggers
• Use phased purchases or dollar-cost averaging to manage volatility and expected corrections while weighing the 0.35% expense ratio against broader, cheaper index alternatives