SOXX - SOXX: AI-Driven Demand May Be Relegated To A Select Few

2023-09-27 07:30:20 ET

Summary

• Consensus price targets do not align with long-term EPS growth rates in the semiconductor sector.
• 60% of companies in the sector may continue to have weak earnings in YE24.
• A valuation of 21x PE in YE24 is not very attractive considering a 9% EPS compound annual growth rate.

Overview

The iShares Semiconductor ETF ( SOXX ) provides broad exposure to the Semiconductor sector including chip designers, manufacturers, and equipment supply chain. Performance has been very strong, up 500% in the last 10 yrs, outpacing the NASDAQ (NXP) and almost identical to peer ETFs.

As many investors may be aware, semiconductors are key building blocks in the digital age. However, demand is cyclical, generally ebbing and increasing with global GDP. There are many chip applications from PC and smartphones, blockchain data crunchers, graphics chipsets to more simplistic sensors that have impacted the auto makers and now AI data processing. Most companies are application focused, a few companies, such as Nvidia ( NVDA ) have been able to weave into new applications. Then there are the manufacturers such as TSM that build chips for everyone else and finally the equipment suppliers such as ASML and LRCX .

This cyclicality is evident in YE23, with many stocks posting earnings declines. Some appear to rebound rapidly in YE24, and others look quite weak for longer.

I calculated that the ETF weighted portfolio has 9% EPS CAGR (YE22-YE25) with a PE of 21x and upside potential of 24% on consensus estimates.

Looking Under the Hood

I analyzed the 30 stocks that make up the ETF. Using consensus data; EPS, Price target, Revenue, and Net debt. The weighted upside on consensus price targets looks quite appealing at 24% to YE24. However, the EPS CAGR of 9% (YE22-YE25) is low in my view. This slow growth is a function of high pandemic results followed by a cyclical decline post pandemic.

Significant Earnings Disparity Projected

Consensus year-over-year (YoY) EPS growth rates highlight significant disparity among the companies. The majority have declining EPS estimates in YE23 which drives a weighted EPS growth rate of 1%. This disparity follows into YE24 estimates with EPS slated to grow 27%. However, this is an optical illusion, most companies will not yet return to YE22 absolute earnings levels. Notably weak stocks are Micron (MU) and Qualcomm ( QCOM ).

Structural Earnings Weakness

To further illustrate this structural earnings weakness, I rebased EPS to 100 in YE22. As an example, Lattice ( LSCC ) estimated EPS grows to 119 in YE23, 139 in YE24 160 by YE25.

As can be seen in the table below, I calculated that only 5 stocks have growth estimated in YE23 i.e., above 100. While in YE24, just 11 see positive earnings growth vs YE22. This erratic and/or cyclical earnings data suggests that the bulk of the semiconductor sector "Over Earned" during the pandemic and many are seeing downturns through YE25 i.e., any stock not above 100.

Valuation

The PEG (PE/EPS Growth) metrics should factor in the volatility of EPS consensus estimates during the YE22 to YE25 timeframe. Many stocks appear to experience extended earnings contraction in absolute terms from YE22 to YE25. This is not captured in the YoY EPS growth rates and distorts the earnings power of many companies.

For example, Intel ( INTC ) is expected to see a 68% EPS decline in YE23 and then a 190% gain in YE24 but in absolute term EPS is still below YE22 levels. Is Intel cheap at 19x PE if long-term growth is estimated at 8%? I don't think so.

Using longer-term EPS (consensus) CAGR from YE22 to YE25 I calculated a weighted PEG of 2.4x, EPS growth of 9% vs YE24 PE of 21x. While this is not terrible vs the SP500 (SPX) or even Nasdaq it seems to me that investors can find better options.

Conclusion

In my view, this ETF is unappealing from a growth and valuation standpoint. The YoY EPS growth estimates are an optical illusion in many cases and represent recuperation vs actual growth as companies deal with a cyclical downturn. AI-driven demand seems to be relegated to a few. In addition, the analyst price targets and valuation do not seem to coincide with longer term earnings power of this sector.

For further details see: