Michael Burry was among a small minority of investors who predicted the housing market collapse that preceded the Great Recession. The S&P 500 fell by more than 50% as those events unfolded, erasing trillions of dollars in wealth. But Burry turned a tidy profit. He pocketed $100 million and made another $700 million for hedge fund clients by betting against subprime mortgage bonds. Those events inspired the 2015 film The Big Short.

Burry turned his attention to soaring semiconductor stocks in the third quarter. His hedge fund, Scion Asset Management, purchased $47 million in put options -- contracts that confer the right to sell a security at a specific price within a certain time frame -- on the iShares Semiconductor ETF (SOXX 2.11%). That index fund tracks 30 chipmakers and semiconductor equipment suppliers, including Nvidia (NVDA 6.18%) and Advanced Micro Devices (AMD 2.37%).

The iShares Semiconductor ETF has gained 47% year to date, more than doubling the return of the S&P 500. Those gains come amid a flurry of enthusiasm surrounding artificial intelligence (AI), but Burry clearly sees substantial downside. The put options Scion purchased account for nearly half of its portfolio.

Should investors avoid semiconductor stocks?

Some popular semiconductor stocks trade at pricey valuations

The largest holdings in the iShares Semiconductor ETF are Nvidia, AMD, Broadcom (AVGO 3.84%), and Intel (INTC -9.20%). That quartet of companies accounts for 33% of the index fund by weighted exposure, and all four stocks have crushed the S&P 500 this year, with returns ranging from roughly 65% for Intel to roughly 220% for Nvidia.

Each company retains reasonable growth prospects as the artificial intelligence (AI) boom continues. But all four stocks now trade at elevated valuations, as detailed below:

  • Nvidia invented the graphics processing unit (GPU), chips that have become the gold standard in accelerating complex data center workloads like 3D graphics and AI. The company holds more than 90% market share in workstation graphics processors and north of 80% market share in AI computing. The stock currently trades at 25.2 times sales, a premium to the three-year average of 23.3 times sales.
  • AMD has steadily gained ground in the central processing unit (CPU) space in recent years, and that trend continued in the third quarter. Its market share now sits at 20% in personal computers and 23% in data center servers. AMD trails Nvidia by a wide margin in data center accelerators, but it recently acquired AI start-up Nod.ai to boost its software capabilities, and its latest AI chip is on pace to reach $1 billion in sales faster than any product in AMD history. The stock currently trades at 8.6 times sales, a premium to the three-year average of 7.9 times sales.
  • Broadcom is the market leader in application-specific integrated circuits, a type of chip built for specific use cases. For instance, Alphabet and Arista Networks both use Broadcom chips in their products, and both should benefit from growing demand for AI. To that end, several analysts believe Broadcom could be the second-biggest winner from the generative AI boom behind Nvidia. The stock currently trades at 11.1 times sales, a premium to the three-year average of 9.2 times sales.
  • Intel still dominates the CPU space, with more than 62% market share in the third quarter, and its leadership position spans both personal computers and data center servers. Intel hopes to maintain its dominance in both markets by tailoring its CPUs to AI uses cases, and the company also hopes to make a dent in the data center accelerator market with its Gaudi processors. The stock currently trades at 3.3 times sales, a premium to the three-year average of 2.5 times sales.

Investors should tread cautiously

Grand View Research says the market for AI hardware, software, and services will grow at 37% annually through 2030. Bloomberg Intelligence thinks the generative AI hardware market will compound at 33% annually through 2032. And Morgan Stanley believes recent breakthroughs in AI could unlock a $6 trillion opportunity across industries like e-commerce, digital advertising, and cloud computing.

Suffice it to say, AI is projected to be a booming industry, and some semiconductor stocks will undoubtedly benefit. Investors should exercise caution in the current environment given the elevated valuations. But I see no reason to avoid chipmakers or equipment suppliers entirely, despite Burry's bearish bet.

Investors should always consider valuation in the context of future growth prospects, but they should avoid prioritizing valuation over quality. To quote Warren Buffett, "It's far better to buy a wonderful company at a fair price than a fair company at a wonderful price."

In my estimation, Nvidia is a wonderful company. Analysts generally think the chipmaker will maintain its dominance in AI computing, and that should drive strong sales growth for years to come. Indeed, Morgan Stanley says Nvidia could grow sales at 30% annually through 2027, and Morningstar says the company could grow sales at 22% annually through 2033. Those forecasts make its current valuation of 25.2 times sales seem "fair-ish," though certainly not cheap.