One of the Biggest Chip ETFs Has Averaged 14% a Year Since 2001 and Just Made 118% in Twelve Months

The iShares Semiconductor ETF (NASDAQ: SOXX) has returned about 118% over the past 12 months, dividends included. The same fund has averaged 14.2% a year since it launched on July 10, 2001.

Set those two numbers side by side, and the past year starts to look surreal. At the fund's long-run pace, a 118% gain takes about six years to accumulate. Semiconductor investors just collected it in one, as spending on the graphics processing units (GPUs) and infrastructure behind artificial intelligence (AI) boomed.

The fund holds about $43 billion in assets, making it one of the biggest semiconductor funds investors can buy. So plenty of people now own a piece of this run -- and plenty more are likely wondering whether to chase it.

What has happened after this fund's other enormous years?

Image source: Getty Images.

Only two precedents

Work through the fund's monthly price history, and only two earlier stretches come anywhere near a year like this one. The 12 months through January 2004 delivered about 97%, as chip stocks rebounded violently from the dot-com bust. And the 12 months through March 2021 produced about 109%, powered by the pandemic's electronics boom and the first wave of chip shortages.

That's the whole list -- and neither quite matches this one. The trailing-year figure reached about 170% at the end of June, far beyond anything in the fund's 25 years, before the recent pullback trimmed it to about 118%.

Year two cooled both times

What followed those two earlier runs is where the record gets uncomfortable. Both episodes disappointed in the year that followed, then split completely.

After the early 2004 peak, the fund fell about 23% over the next year. Three years on, it still sat below where the run ended. And five years on, it was down more than half -- though that stretch admittedly ended in January 2009, in the depths of the financial crisis. An investor who bought at the end of that first run waited about two years just to get back to even.

The 2021 episode started better, but not by much. The next 12 months returned about 12%, well below the fund's long-run average, and 2022 then took the fund down 35% before it stabilized.

But the five-year outcome could hardly have been more different. The fund returned about 142% in total over that span, or about 19% a year, because the AI build-out arrived and handed the industry a bigger demand wave than the pandemic ever created.

In other words, the two precedents agree about the year that follows a run like this. It fell short of the long-run average both times. They disagree completely about the five years after. And what separated them wasn't the size of the run. It was the demand: a new wave showed up after 2021, and nothing comparable did after 2004.

Which history applies now?

Today's version of that test is easy to describe. The fund's holdings trade at about 67 times earnings.

Its three biggest positions are Nvidia at about 9% of assets, Advanced Micro Devices at 8.2%, and Broadcom at 8.2%, and all three are riding the same AI spending cycle. About a quarter of the fund sits in those three names.

In short, this is not a diversified bet that semiconductors matter -- it's a concentrated bet that AI infrastructure spending keeps compounding, at prices that already assume it does.

It might. After all, the 2021 buyers who looked wrong in late 2022 were rescued by exactly that kind of surprise. But at 67 times earnings, after a 118% year, the fund is priced as if the demand wave keeps building for years to come. And after both of its earlier big runs, the next year came in well below the fund's long-run average.

The fund itself is a fine vehicle, with 30 holdings and a 0.33% expense ratio. The uncomfortable part is the starting valuation. A weak year two isn't guaranteed, and the AI cycle may prove more durable than the two booms before it. But I think the 14.2% long-run average is the better number to plan around. Much of what the past year delivered above that pace arguably belongs to years that haven't happened yet.

Daniel Sparks and his clients do not have positions in any of the stocks mentioned. The Motley Fool has positions in and recommends Advanced Micro Devices, Broadcom, Nvidia, and iShares Trust-iShares Semiconductor ETF. The Motley Fool has a disclosure policy.