Most development in artificial intelligence (AI) happens inside data centers, which are filled with thousands of purpose-designed chips from suppliers like Nvidia, Advanced Micro Devices, and Broadcom. In order for AI models to continue improving, they require more computing power so they can process higher volumes of data at faster speeds.
In fact, Nvidia CEO Jensen Huang thinks the latest reasoning models, which spend more time "thinking" before rendering outputs, could drive an AI infrastructure spending boom worth $4 trillion by 2030. If that's the case, it might be a great idea for investors to own a slice of the semiconductor industry.
The iShares Semiconductor ETF (SOXX +0.33%) is an exchange-traded fund (ETF) that invests exclusively in companies that supply chips and other computing hardware components. It manages a highly concentrated portfolio of just 30 stocks, and Advanced Micro Devices (AMD), Nvidia, and Broadcom are among its top holdings.
Here's how the ETF could turn a $250,000 investment into $1 million over the long term. But don't worry -- investors with any starting balance can earn a fourfold return if this scenario plays out.
Image source: Getty Images.
A complete portfolio of AI hardware stocks
The iShares Semiconductor ETF invests in companies that design, manufacture, and distribute chips and components, but especially those that stand to benefit from megatrends like AI. Therefore, its top three holdings should come as no surprise:
|
Stock |
iShares ETF Portfolio Weighting |
|---|---|
|
1. AMD |
9.77% |
|
2. Nvidia |
8.57% |
|
3. Broadcom |
8.17% |
Data source: iShares. Portfolio weightings are accurate as of Aug. 29, 2025, and are subject to change.
AMD's newest lineup of graphics processing units (GPUs), called the MI350 series, is being adopted by leading data center operators like Oracle. The company will up the ante in 2026 with the MI400, which could be 10 times more powerful. But AMD's AI opportunity transcends the data center, because it's also a leading supplier of AI chips for personal computers, which could be a major growth market in the future.
Despite the competitive threat from AMD, Nvidia still has a decisive edge in data centers thanks to its latest Blackwell Ultra GPUs, which are the gold standard for AI reasoning models. These chips are 50 times more powerful than the company's original Hopper AI GPUs, which led the industry in 2023. Nvidia generates around 12 times more data center revenue than AMD, which highlights its incredible market share.
NASDAQ: SOXX
Key Data Points
Broadcom designs AI accelerators (a type of data center chip) for hyperscale cloud providers like Alphabet's Google Cloud. These chips can be customized to suit specific workloads, so they offer more flexibility than traditional GPUs. But Broadcom is also a leading supplier of data center networking equipment, like its latest Tomahawk Ultra Ethernet switch, which provides industry-leading low latency and high throughput to drive faster processing speeds in AI workloads.
Since the beginning of 2023, which is when the AI boom started gathering momentum, the above three stocks have delivered an average return of 550%, with each of them crushing the S&P 500:
And the iShares ETF holds a number of other top AI hardware stocks, including Micron Technology, which supplies some of the industry's best storage and memory chips, and Taiwan Semiconductor Manufacturing, which fabricates many of the AI chips designed by companies like Nvidia and AMD.
Turning $250,000 into $1 million
The iShares Semiconductor ETF has delivered a compound annual return of 11.4% since its inception in 2001, comfortably outpacing the average annual gain of 8.5% in the S&P 500 over the same period.
And the fund generated an accelerated compound annual gain of 24.1% over the last 10 years, thanks to the proliferation of technologies like cloud computing, enterprise software, and AI, which continue to drive explosive demand for chips.
Here's how long it could take the iShares ETF to turn an investment of $250,000 into $1 million, based on three different compound annual returns.
|
Initial Investment |
Compound Annual Return |
Time To Reach $1 Million |
|---|---|---|
|
$250,000 |
11.4% |
13 Years |
|
$250,000 |
17.7% |
9 Years |
|
$250,000 |
24.1% |
7 Years |
Calculations by author.
A continued annual return of 24.1% over the long term would be a very big ask for this ETF -- or any ETF -- because there is no such thing as infinite growth. Nvidia, for instance, is already the world's largest company, and its revenue growth is starting to decelerate because there are only so many new customers capable of buying enough chips to move the needle.
With that said, the semiconductor industry as a whole can still grow materially from here if AI infrastructure spending hits $4 trillion by 2030, as Jensen Huang predicts, so the iShares ETF is likely still a great buy. Adding it to a diversified portfolio of other ETFs and individual stocks can protect against some of the volatility that is typical in high-growth industries like AI.
