Update: semi-conductor and hyper-scaler stocks, plus SpaceX

 U.S. and Asian stocks remain under pressure, especially in the highly-volatile semi-conductor sector and across the entire A.I. value chain.  I wanted to take a quick look at the semi-conductor companies, the Mag 7 cohort (four hyper-scalers), and SpaceX.  The tables in this update contain data as of the close of business on Friday, July 24^th.  For the record, most of the semi-conductor stocks closed yesterday lower than on Friday (and are down today sharply in Asia), and the Mag 7 stocks generally closed higher.

Semi-conductors: one component of the A.I. value-chain

I wrote about the semi-conductor companies in my blog on June 25 here.   Since the end of June, the IDM stock prices have fallen between 20% and 34%, although all have realised significant gains YtD.  Foundries are down 15% to 20%, and fabricators/fabless companies are down 10% to 20%, both also since the end of June.  The sector remains highly volatile, as we saw yesterday again and this morning, with the chip-heavy KOSPI (Korea) down over 10% and the TWSE (Tawain) down 4.6%.

The table below contains return data and some key valuation metrics for many of the players in the semi-conductor value-chain.   Keep in mind that the data is as of Friday’s close (July 24), and most semi-conductor stocks have moved sharply lower since then.

 Memory IDMs (companies on the left) look very cheap on traditional valuation metrics, so cheap that it makes me wonder why?  Of course, they are all delivering exceptional top- and bottom-line growth and have juicy margins, reflecting strong demand from hyper-scalers that can’t buy memory chips for data centres fast enough.  The exception is INTC, a transitioning company and one in which the U.S. government has an ownership stake.

Looking at the valuation metrics and focusing on IDMs, I would actually feel more comfortable shorting INTC at this level than being long SKHY, Samsung or MU, at least until the volatility settles down.  As my readers know, I like NVDA as a fabricator, and ASML (which I forgot to include), a monopolistic semi-conductor equipment company (makes lithography equipment).  And as I mentioned in the recent article on my blog, I would consider both TSM and SOXX on weakness.  I am starting to build a position in TSM now, as the stock is getting battered.  It won’t be a runner like the memory chip companies, but it is at the nucleus of chip manufacturing for fabless players like NVDA, AMD and Broadcom.

I would be remiss if I were not to mention the IPO of Chinese IDM CXMT on Monday, which was listed on the Shanghai Stock Exchange.  As you can see in the graph below from the #FT, CXMT is smaller than its peers. 

In spite of its size, CXMT is apparently well positioned for the huge growth in memory chips in China and globally.  The shares soared 466% on Monday, becoming the most valuable company in China.  You can read more in the #FT in an article published yesterday (free for FT subscribers): “Chinese chip champion CXMT soars 466% in market debut”. 

Hyper-scalers: another component of the A.I value chain

Four of the seven Mag 7 companies are hyper-scalers, spending like mad to meet the insatiable demand for data centres as A.I. explodes.  Another one of the companies is a fabricator (NVDA, included in both tables), the sixth one is mostly a consumer company (AAPL), and the last one (TSLA) is an automotive “want to be an A.I., robotic and autonomous driving champion” company.

Interestingly on Monday, most of these companies registered gains while the semi-conductor stocks were losing ground.  My faithful readers know I own five of these companies – MSFT, AAPL, AMZN, GOOG and NVDA.  As with semi-conductors and INTC, the company that most stands out be me in this table is also a short – TSLA.  Investors have been hammered owning TSLA this year, and the stock remains significantly over-valued in my opinion based on aspirational dreams of Elon Musk and his army of faithful followers.  I’m not sure why.  Having said this, I also think AAPL has gotten really expensive.  I suppose it has been easier to bet on AAPL amongst the Mag 7 than the four hyper-scalers, in that AAPL is not spending like a drunken sailor on data centres.  But sporting a nearly 35x forward P/E ratio and the highest PEG ratio (2.68x) amongst its peers is more than concerning to me.  I have been saying this for a long time and selling into strength accordingly.  However, I started lightening up on AAPL, but this started over $100 ago so I left plenty of money on the table.  The shares have been the best performing Mag 7 stock this year, and the company has now overtaken NVDA as the largest company in the world with a market cap approaching $5 trillion. 

I have added clips of MSFT and GOOG to my portfolio on weakness in the last week, but have already added clips to MSFT on weakness several times this year.  So far, it has been like catching a falling knife, as the stock is struggling to get and stay back above $400/share.  I remain a believer though, recognising that the software component of its business will remain under pressure from the “other side” of A.I. (similar to all software companies that are threatened by A.I.).  However, the company has weathered worst, remains a leading cloud company, and is one of only two companies with a coveted AAA rating (higher rating than the U.S.).

This week we have earnings from AAPL, AMZN, META and MSFT.  These will be very influential announcements as far as the market at large, with investors keen to understand the outlook for the hyper-scalers in the coming quarters, following GOOG’s earnings last week.

SpaceX: it’s a dog so far!

I wrote about the SpaceX IPO on June 4 here.   Here’s the chart of the stock’s performance since the IPO.

I suppose it wouldn’t be nice to say “I told you so”, but it was rather obvious that the shares were massively over-valued at the IPO for similar reasons that TSLA has defied gravity for years.  Moreover, the lead underwriters pumped the stock like it was the most amazing company ever.  That lasted a few days, and then down to Earth SPCX has come. 

I think the stock has much further to fall, and wouldn’t even consider it at half its current price.  On one hand, there is very limited float so the shares will occasionally find support among the diehards.  On the other hand, insider lock-ups are running off soon, and this will massively increase the selling pressure on the stock.  At the end of the day, investors have to ask if the lead banks did this company a dis-service, falling into the honey trap of Elon Musk and his aspirational dreams.  Certainly, investors cannot be happy.  My advice is simple – unless you believe the dream and really embrace the hype around A.I. (which recall is meant to be the most substantial component of the heady valuation, many times more than rocket ships and Starlink), then I wouldn’t touch this stock with a 10-foot pole.