Another way to play AI stocks: These have been the most efficient chip makers

Too expensive?

First, let’s take a look at valuations. Here is Nvidia’s year-to-date price chart through Friday:

The stock popped 24% on May 25 after Nvidia projected a 50% sequential increase in sales for the current quarter, with that expectation tied to the adoption of the company’s graphics processors for AI computing in data centers.

Now let’s take a look at Nvidia’s valuations compared with those of the S&P 500
SPX,
-0.17%
and the iShares Semiconductor ETF
SOXX,
+0.99%,
which tracks the PHLX Semiconductor Index
SOXX,
+0.99%
of 30 manufacturers of computer chips or related hardware.

Here are forward price-to-earnings ratios, based on weighted rolling consensus estimates for 12 months among analysts polled by FactSet:

And now, forward price-to-sales ratios:

Some points about the valuation charts:

• Nvidia’s forward P/E of 50.7 compares with valuations of 18.8 for the S&P 500 and 22.5 for SOXX.

• SOXX trades at less than half of Nvidia’s P/E, even though Nvidia makes up 12.3% of the exchange-traded fund’s portfolio. That highest weighting is followed by Broadcom Inc.
  AVGO,
  +0.64%
  at 9.3% of the portfolio and a P/E of 18.9 and Advanced Micro Devices Inc.
  AMD,
  -1.19%
  at 7.2% of the portfolio and a P/E of 32.1.

• Nvidia’s forward P/E of 50.7 is only 2.2 times its forward price-to-sales ratio of 22.9, emphasizing how profitable its business is expected to be. In comparison, the S&P 500’s P/E of 18.8 is more than eight times its price/sales ratio of 2.3. For SOXX, the P/E of 22.5 is 4.4 times its price/sales ratio of 5.1.

We might look back in a few years and conclude that late June of 2023 was still a good time to buy shares of Nvidia, because of its high profitability and incredible growth trend. But clairvoyance is in short supply.

SOXX can be a reasonable way to play the AI/semiconductor trend while avoiding having all your eggs in the Nvidia basket.

Or you may target AI itself, beyond chip makers, by investing in focused ETFs that take various approaches, including these three:

• The Global X Robotics & Artificial Intelligence ETF
  BOTZ,
  -0.20%
   holds 43 stocks. It tracks an index of companies listed in developed markets. The companies are expected to benefit from the increased utilization of robotics and artificial intelligence. The fund is weighted by market capitalization; its largest holding is Nvidia, which makes up 12.4% of the portfolio, according to FactSet. It is the largest AI ETF listed here $2.4 billion in assets under management. It was established September 2016.

• The iShares Robotics and Artificial Intelligence Multisector ETF
  IRBO,
  -0.03%
   holds 116 stocks that are equal-weighted, as it tracks a global index of companies that derive at least 50% of revenue from robotics or AI, or have significant exposure to related industries. This ETF has $408 million in assets and it was launched in June 2018.

• The $361 million First Trust Nasdaq Artificial Intelligence & Robotics ETF
  ROBT,
  +0.07%
   has 107 stocks in its portfolio, with a modified weighting based on how directly they are involved in AI or Robotics. It was established in February 2018.

Semiconductors — ROIC screen

Getting back to chip makers and companies that manufacture the equipment they use or provide related services, we began with the SOXX 30 and then added the 30 additional companies in the S&P 1500 Composite Index in the semiconductor industry, as determined by FactSet, or in the “Semiconductors and Semiconductor Equipment” Global Industry Classification Standard (GICS) group. (The S&P 1500 Composite is made up of the S&P 500, the S&P 400 Mid Cap Index
MID,
+1.03%
 and the S&P Small Cap 600 Index 
SML,
+1.05%.
)

We then looked back at returns on invested capital for the companies’ past 20 reported fiscal quarters, to calculate five-year averages.

FactSet defines a company’s return on invested capital (ROIC) as earnings divided by the sum of the carrying value (not the market value) of a company’s common stock, preferred stock, long-term debt and capitalized lease obligations.

ROIC is most useful within specific industry groups for a comparison of how well management teams allocate investors’ money, because some industries naturally have higher profit margins than others. It isn’t a perfect measure and it is important to do your own research if a ROIC number makes a company appear attractive. For example, a company’s equity capital on its balance sheet may actually be negative, and this may not be a problem if it is profitable with good cash flow. But it can distort ROIC.

It turns out that among our 60 semiconductor companies, there are four for which less than 20 quarters of ROIC data are available.

Among the 56 remaining companies, there are 20 with an average ROIC of 15% (if we round up) or more over the past 20 quarters:

Click on the tickers for more about each company, ETF or index.

Click here for Tomi Kilgore’s detailed guide to the wealth of information available for free on the MarketWatch quote page.

There is no forward P/E ratio for Micron Technology Inc.
MU,
+0.87%
because the sum of analysts’ consensus earnings-per-share estimates for the company’s next four reported fiscal quarters is negative.

There is some correlation between high ROIC and total returns. Out of these 20 companies, 16 have beaten the S&P 500’s 72% total return over the past five years, with dividends reinvested. Meanwhile, 11 of them have beaten the 176% five-year return for SOXX.

It might surprise you to see Intel Corp.
INTC,
+0.94%
on the list, with a 16.55% average ROIC. This is the only company that made the top 20 with a negative five-year total return. Intel is undergoing an attempted multiyear comeback, including the creation of a large chip foundry business in the U.S. and other Western countries.

Keep in mind that ROIC is a useful tool for an initial indicator of how efficient a company has been at making use of investors’ money. It is a first step that can point to companies worthy of further research as part of your effort to find companies likely to remain competitive providing goods and services over the next decade, at least.

Related: If you own these stock index funds, you’re already invested in AI. But there may be better strategies.