SMH: Benefitting From Specialized AI Chips Demand, But Is Also At Risk (NASDAQ:SMH)

The VanEck Semiconductor ETF (NASDAQ:SMH) is at ground zero of the spectacular market rally in chip stocks benefiting from the artificial intelligence boom. The rise of generative AI and large language models has set off a wave of demand for the next generation of chips specialized to handle complex computations.

Recognizing that NVIDIA Corp (NVDA) plays a key role with its AI-optimized GPUs, the entire industry is capturing strong growth across chip designers, manufacturing leaders, and service providers. Indeed, SMH has climbed more than 67% over the past year, including a 16% gain to start 2024.

In our view, SMH is a good option to capture these high-level trends driving chip stocks. At the same time, the caution here is that SMH is distinctly concentrated among the top holdings, with NVDA alone representing nearly 25% of the current portfolio. While shareholders likely won’t complain given the stellar recent performance, there are some important risk implications to consider with the potential for higher volatility in the future.

What Is the SMH ETF?

SMH is intended to track the “MVIS US Listed Semiconductor 25 Index”. As the name implies, the strategy considers the 25 largest U.S.-listed companies that derive at least 50% of their revenues from the production of semiconductors and related equipment.

Through a free-float weighting methodology, there is a 20% weighting cap that is rebalanced quarterly, with the next review coming in March. We bring this up because given the nearly 50% rally in NVDA year-to-date, the stock’s current composition in the fund at 24% will likely be marginally reduced at the next rebalancing.

Taiwan Semiconductor Manufacturing Co (TSM) with a 10% weighting is the next largest position in the fund. Keep in mind that TSM is the primary manufacturer of Nvidia’s GPUs, which highlights both the importance of NVDA to the industry, but also the higher indirect exposure to NVDA within SMH above its 24%.

Down the list of holdings, other high-profile names are well-represented by Broadcom Inc. (AVGO) and Advanced Micro Devices (AMD) with a 6% weighting, followed by ASML Holding N.V. (ASML) at about 5%.

While AI is a major theme in the industry, there are other drivers at play, including chips for infrastructure and networking, general computing, mobile, and even automotive applications.

The chart below shows that, except for a few laggards, the strength in chip stocks has been widespread among large-caps in the early part of 2024.

SMH NVIDIA Concentration

We mentioned the high portfolio concentration. For context, the top 10 holdings in SMH comprise approximately 72% of the total fund. Typically, we’d say that a 24% allocation to a single stock nearly defeats the purpose of an ETF where diversification is often an important consideration. For someone very bullish on NVDA, buying the stock directly, maybe through a smaller position, can make more sense.

On the other hand, the angle that could work with SMH is precisely that “NVDA+” profile, where investors can gain significant exposure to arguably the most important chip stock in the market while supplementing that position with a broader basket of industry leaders. In other words, an allocation to SMH can be seen as a moderately bullish view on NVDA that still limits company-specific risk.

The other option is through alternative ETFs like the iShares Semiconductor ETF (SOXX) where NVIDIA has a less dominating 10% weighting in the fund, which is overall less concentrated with 30 current holdings. There is also the smaller SPDR S&P Semiconductor ETF (XSD) which stands out with its equal-weighting methodology, leaving NVDA with just a 3.8% weighting.

There’s no right or wrong answer here, in terms of which strategy is optimal or best over the long term, but simply that investors should be aware of these weighting differences between benchmark semiconductor ETFs.

What we do know is that SMH is the largest fund in terms of AUM with $15 billion, compared to SOXX at $12 billion, and XDS at $1.5 billion. Notably, all three here feature a 0.35% expense ratio.

What’s Next For SMH?

There are plenty of reasons to believe the AI tailwind for chip stocks is here to stay as a secular growth driver. The idea here is that with AI applications extending across several sectors, the demand for chips should increase dramatically as incremental to the typical computing replacement cycle.

Earnings reports this quarter from companies like Broadcom and ASML have included positive guidance for the year ahead, suggesting good visibility for the operational momentum to continue over the next several quarters.

That said, given the strong stock price gains over the past year, a consequence of that has been rising valuation multiples. While far from anything we would consider resembling a “bubble”, metrics like the price-to-cash flow ratio or price-to-earnings multiple over the trailing twelve months have climbed over the last several quarters.

At the index level, SMH price-to-earnings multiple of 34x, which is well above the sub-16x range at the start of 2022. With some room for earnings to grow into this multiple, the key is for companies to keep up with expectations.

Maybe the biggest risk for chip stocks is on the macro side. New questions on the path of inflation and the next steps in Fed policy add a layer of risk to the investing backdrop.

Looking ahead, all eyes will be on Nvidia, set to report its Q4 results on February 21st. The company is forecast to revenue growth of 235% and EPS that is up 416% higher than the period last year. With expectations already so high, evident by the recent strength of the stock, the challenge we see will be for the company to once again surprise the market to the upside, which would be important to set the tone for the rest of the industry.

By this measure, NVDA earnings are a high-risk event for the SMH ETF, given its outsized weighting in the fund and ability to influence sentiment toward the broader portfolio. The risk is that disappointing numbers or weaker guidance forces a reassessment of the earning momentum.

Final Thoughts

SMH’s large position in Nvidia can be seen as both part of the fund’s attraction or a weak point of its composition structure, depending on the perspective.

Investors can use the fund as a long-term strategic holding in the context of a more diversified portfolio, or as a tactical trading instrument for targeted exposure to key chip stocks.

We set a tactical sell rating on the fund, with an expectation that the next big move is lower. From the more uncertain macro setup to the appearance that chip stocks are technically overbought, patient investors be able to find a more attractive lower entry point down the line as a new buying opportunity.