BrokerHiveXBrokerHiveX
Global AI chip ETF asset size exceeds US$200 billion, semiconductor investment boom escalates
forex5 months before
Summary：According to Resonance Analytics, an international ETF data platform, as of July 2025, the total management scale (AUM) of the world's seven mainstream AI chip theme ETFs exceeded US$200 billion for the first time, a surge of 85% over the same period last year. Benefiting from government subsidies, corporate computing power expansion and the blowout of generative AI applications, the semiconductor industry chain has ushered in a new round of capital competition. Experts warn that high valuations and geopolitical supply chain risks may become the main triggers for the next stage of fluctuations. According to Resonance Analytics, an international ETF data platform, as of July 2025, the total management scale (AUM) of the world's seven mainstream AI chip theme ETFs exceeded US$200 billion for the first time, a surge of 85% over the same period last year. Benefiting from government subsidies, corporate computing power expansion and the blowout of generative AI applications, the semiconductor industry chain has ushered in a new round of capital competition. Experts warn that high valuations and geopolitical supply chain risks may become the main triggers for the next stage of fluctuations.
Contributor
Andrew Li , Senior Reporter, Technology and Capital Markets
1. ETF scale rises exponentially
VanEck Semiconductor AI Leaders ETF (SMAL) : AUM is US$72 billion, up 52% from the beginning of the year.
BlackRock Global AI Chip Innovators ETF (AICI) : AUM $61 billion, with approximately 16 million new shares.
ARK Next-Gen Compute ETF (ARKC) : AUM $31 billion, thanks to U.S. IRA (Inflation Reduction Act) tax credits.
In the past 12 months, institutional funds have net inflows of more than $82 billion; 70% of which came from US and European pension funds and sovereign wealth funds, and 30% from Asia-Pacific hedge funds. Resonance Analytics predicts that if the US and EU continue to promote local chip incentive policies, the AI chip ETF market size is expected to exceed $300 billion in 2026.
2. Resonance between policy drive and industry catalysis
| Drivers | Specific policies/events | Direct impact |
|---|---|---|
| Government subsidies | The US "Chips and Science Act" and the EU "Chips Act" allocate a total of $85 billion | Driving capital expenditure of local wafer fabs |
| AI computing power requirements | OpenAI GPT-6, Google Gemini Ultra training parameters exploded | High-end HBM memory and GPUs are in short supply |
| Cloud vendor centralized procurement | Amazon, Microsoft, and Tencent announce GPU purchases of more than $10 billion from 2025 to 2027 | Improve order visibility for equipment manufacturers |
3. Valuation and risk: bubble or starting point?
Higher valuations : The price-to-sales ratios (P/S) of star chip stocks generally rose to 25-30 times, much higher than the historical average.
Supply chain constraints : Geopolitical frictions have led to tight production capacity for high-end EUV equipment and advanced packaging; once export controls are tightened, the profits of ETF heavyweights may be squeezed.
Technological iteration : The technological watershed of AI-specific chips (ASICs) replacing general-purpose GPUs will influence the industry landscape in the next 3-5 years.
Morgan Stanley warned that if orders slow down in the short term, ETF volatility may be significantly amplified; it recommends investors to keep their positions flexible and allocate safe-haven options.
4. Derivative Trends: Hash Power REIT and Hash Power Bonds
Computing power data center REIT : Blackstone and Digital Realty packaged GPU cluster assets and listed them, with an expected annualized return of 8-10%.
Compute Bonds : NVIDIA plans to issue $1 billion in AAA-rated bonds secured by future GPU cash flows to provide low-cost funds for supply chain expansion.
5. Long-term Outlook: Resonance among National Policy, Industry and Capital
The industry generally expects that global AI chip investment will remain high in the next five years. Andrew Li pointed out: "From policy subsidies to ETF allocation in the capital market, to computing power REITs and bonds, AI semiconductors have formed a full-stack financial product matrix. The real test is whether the technology cycle and geopolitical risks can enable capital to obtain sustainable returns."
⚠️Risk Warning and Disclaimer
BrokerHivex is a financial media platform that displays information from the public internet or user-uploaded content. BrokerHivex does not support any trading platform or instrument. We are not responsible for any trading disputes or losses arising from the use of this information. Please note that the information displayed on the platform may be delayed, and users should independently verify its accuracy.
