# Why Chip ETFs Are Quietly Winning the AI Race Right Now
Here's the thing: instead of betting on which AI company will "win," smart money is betting on the picks and shovels — the semiconductors that power everything.
The VanEck Semiconductor ETF (SMH) holds 25 of the biggest chip manufacturers. In 2025 alone, it's up 38% vs 17% for Nasdaq. Not bad for a "boring infrastructure play."
Why? The global AI market was $279B in 2024 and is projected to hit $3.5T by 2033. That's 31.5% annual growth. Every single AI training run, every data center, every GPU — they all need chips.
Top holdings: - Nvidia (18.5%) — GPUs that train AI models - Taiwan Semiconductor (9.5%) — manufactures chips for everyone - Broadcom (8.1%) — AI accelerators (just partnered with OpenAI for 10GW custom chips) - Micron (6.8%) — memory chips that keep AI running - AMD (6.6%) — CPUs powering data centers
These five stocks make up nearly half the fund. The expense ratio? 0.35% — actually cheap for a focused ETF.
The real play: you don't need to predict which AI startup survives. You just need the infrastructure that *every* AI company needs. That's the difference between picking winners and owning the entire ecosystem.
This page may contain third-party content, which is provided for information purposes only (not representations/warranties) and should not be considered as an endorsement of its views by Gate, nor as financial or professional advice. See Disclaimer for details.
# Why Chip ETFs Are Quietly Winning the AI Race Right Now
Here's the thing: instead of betting on which AI company will "win," smart money is betting on the picks and shovels — the semiconductors that power everything.
The VanEck Semiconductor ETF (SMH) holds 25 of the biggest chip manufacturers. In 2025 alone, it's up 38% vs 17% for Nasdaq. Not bad for a "boring infrastructure play."
Why? The global AI market was $279B in 2024 and is projected to hit $3.5T by 2033. That's 31.5% annual growth. Every single AI training run, every data center, every GPU — they all need chips.
Top holdings:
- Nvidia (18.5%) — GPUs that train AI models
- Taiwan Semiconductor (9.5%) — manufactures chips for everyone
- Broadcom (8.1%) — AI accelerators (just partnered with OpenAI for 10GW custom chips)
- Micron (6.8%) — memory chips that keep AI running
- AMD (6.6%) — CPUs powering data centers
These five stocks make up nearly half the fund. The expense ratio? 0.35% — actually cheap for a focused ETF.
The real play: you don't need to predict which AI startup survives. You just need the infrastructure that *every* AI company needs. That's the difference between picking winners and owning the entire ecosystem.