Modern semiconductor chips are defined by software, but it’s a software market where what you did yesterday does not guarantee success today. A chip’s design is based in software, and most chips are designed around executing a specific kind of software. Hardware has become software.
This has been true for over a decade, ever since “Moore’s Second Law” — the reality that manufacturing equipment costs double roughly every four years — began separating the production of chips from their design.
Everyone loves Moore’s Law, the idea that a chip’s capacity may double every year or two, with no increase in cost. Moore’s Second Law, however, is what drives today’s market, and it isn’t properly understood by investors.
Computer chip manufacturing is highly capital-intensive. Chip manufacturers must find enormous amounts of business to keep plants operating. Relying on their own designs is not good enough. Chip fabrication plants or “fabs” have become like oil refineries, so what matters to their owners is the total market, not the fate of a single design.
The hot chip companies today are “fabless” — they order copies of their designs as sales scale. They may offer only chip designs but the point is they rent factories, they don’t own them. Customers ordering in bulk may be able to make their own tweaks to a design, or they may not. The key is that the design — and the intellectual property behind the design — is what makes a chip company valuable to investors.
Since they don’t have to carry the capital costs of manufacturing, fabless chip companies can be fabulously profitable, if their designs meet a market need and are superior to others, or the intellectual property in them is vital to the final design of a product.
The closer a chip company can get to the customer, in other words, the more vital it can make its software and other intellectual property, and the fatter its profit margins can be. The more reliant it is on others to create downstream demand, the more its margins may be threatened.
This is a good rule of thumb, but remember that all stock performance remains relative to what came before. A company that captures a niche and manufactures chips may still do better than it did in previous years. A company that owns a retail market may not do as well as before, if its market and margins come under threat.
With that, here are three chip stocks that are hot today, and three that are not.
Hot Semiconductor Stocks: Texas Instruments (TI)
The first patents for integrated circuits were given to Robert Noyce, who later co-founded Intel Corporation (NASDAQ:INTC) and to Texas Instruments Incorporated (NASDAQ:TXN) engineer Jack Kilby. Noyce died in 1990 and when the Nobel Committee finally decided to give its Physics prize for the invention, Kilby was the survivor and winner.
While Intel is a Silicon Valley company, Texas Instruments was founded to serve the oil industry, then became a big defense contractor. I once watched the late Andrew Grove, then Intel’s CEO, dance on a stage wearing a clean room outfit to sell his chips for gaming. No TI executive would ever do that. They have a far more conservative, insular corporate culture.
But slow and steady can win the race. Investors who have stayed with the company known as “TI” over the last 10 years have done better than those in Intel. TXN stock is up 129% over the last 10 years, INTC just 47%.
While Intel stayed with microprocessors, TI decided in the 1980s to focus on digital signal processors or DSPs, which turn analog input like music or TV pictures into digital output. TI is still the leader in DSPs, as well as analog chips and embedded chips. These are not the sexy areas of the semiconductor business, but they represented 85% of TI revenue last year.
Embedded chips are becoming a little sexier as the Internet of Things, which uses such chips, gets rolling. Most analysts see this market hitting $3.7 billion by 2020.
Thus, TI has what analysts call a “runway” — long-term easily predictable growth. Revenue growth has been accelerating, up 13% year over year in the first quarter, with profits for embedded circuits up 28%. This has fueled a 12% rise in the stock since the start of the year, and a price to earnings ratio of 21.5, along with a 50 cent per share dividend yielding 2.45%.
There are currently 34 analysts of TI, 14 of whom call it a buy and only 1 of whom thinks you should sell it. Earnings are expected to grow 10% this year. It is not a sexy stock, and has yet to reach the highs it achieved during the dot-com tech bubble in 2000, but it’s closer to those highs than Intel is.
TI, in other words, isn’t a trader’s stock. It’s an investor’s stock. You might even call it a boring investor’s stock. But boring investors tend to get rich carefully. If that’s your style, this is your stock.
Hot Semiconductor Stocks: Nvidia (NVDA)
I like NVIDIA Corporation (NASDAQ:NVDA). I like it very much. I like their products, I like their positioning. CEO Jen-Hsun Huang can call himself Jensen or Jerry or Jo-Jo, but if I met him and his leather jacket on the street I’d still give him a hug or a high-five.
But I’m not putting my investment dollars into Huang’s company right now. The price of NVDA stock has become ridiculous.
Chip companies don’t have price-to-earnings ratios of 54. They’re not worth 13 times sales. They don’t blow past the price target of a bull like our Tim Biggam and just keep going.
Don’t get me wrong. NVIDIA’s first quarter was spectacular. Revenues of $1.94 billion against $1.31 billion a year earlier. Profits of $507 million against $208 million a year ago. NVIDIA’s debt load is dropping, and operating cash flow is skyrocketing.
This is because NVIDIA is in the sweet spot of the market. Its GeForce graphics line is not just where you want to be for gaming, but for autonomous driving. Graphics chips are also transforming the cloud, augmenting the low-cost silicon muscle they now contain with fast twitch muscle to deliver the instantaneous calculation known as deep learning or Artificial Intelligence.
Deep learning, through clouds, means there is a large and growing market for faster chips in data centers. In 10 years, NVDA stock may justify today’s confidence. Self-driving cars will become a thing, although how many and how fast is open to conjecture.
I am perfectly willing to buy NVIDIA stock, and recommend it, at a reasonable valuation. That valuation would be much higher than Intel and its 15 times earnings. It would be better than Qualcomm, Inc. (NASDAQ:QCOM), which sells at 17 times earnings. It would even be higher than Texas Instruments, at 21 times earnings. Just not 54 times earnings. Overpay for anything and you’re looking for trouble.
Hot Semiconductor Stocks: Advanced Micro Devices (AMD)
Bitcoin, Ethereum and other cryptocurrencies are not created by government fiat, but mined by private individuals who search for valid answers to its encryption puzzles. Each answer represents a coin, and each coin then has a value created through an open market ledger of trades.
Fast processors bring answers to the puzzle faster. Cheaper fast processors lower a miner’s costs. Since Advanced Micro Devices, Inc. (NASDAQ:AMD) sells cheap, fast processors, it is naturally going to benefit as mining goes into overdrive.
AMD also has a much-lower market cap than any popular chip stock, at just $12.5 billion. This makes it more volatile, which in turn makes it of greater interest to traders and those who go in and out of stocks regularly.
For investors, however, what matters is long-term momentum, in sales and profits. For the quarter ending in June, AMD announced revenue of $1.22 billion and earnings of 2 cents. Expectations were for revenue of $1.16 billion and flat earnings. Shares popped on the news.
But Bitcoin represents just a small part of AMD’s sales. Game systems and cloud servers are where AMD is making its biggest gains, thanks to its low-power microprocessors and graphics chips.
While anything is possible, I think it more likely that AMD is going through the same rise-and-fall cycle it has been going through for decades. Intel moves slowly, but it moves. It has vast resources with which to design and build new processors that do everything AMD does, and more. This will happen.
When it does, a year or so from now, AMD is not where you will want to be. But AMD shareholders don’t think that far ahead. If you can stand the volatility, if you pay careful attention, and if you don’t fall in love with your AMD investment, you can still make some money here.
Cold Semiconductor Stocks: Intel (INTC)
Intel stunned the technology world on March 13 by agreeing to buy Mobileye for $15.3 billion.
The Mobileye purchase was hailed, but it represents the company’s problems more than its strengths. As the only U.S. microprocessor company with its own fabs, Intel needs enormous markets to justify its capital expenses. It must set plans years ahead, and ends up focusing on its own needs ahead of customers.
Artificial Intelligence uses a concept called deep learning, and that market exists now with cloud, but Intel has yet to announce plans to sell its new deep learning chip, Nervana, into cloud data centers.
It may be telling that the other big buyer of autonomous car technology in the last year was Samsung Electronic KRW5000 (OTCMKTS:SSNLF), which also owns fabs it must fill and paid $8 billion for Harman International Industries.
Intel is trying to replicate what it did in the PC industry, achieving early dominance through partnerships like those it had back in the day with International Business Machines Corp. (NYSE:IBM) and Microsoft Corporation (NASDAQ:MSFT). Intel also managed to do this in the cloud industry, extending the x86 architecture as a cheap solution for building scaled services early this decade.
But Intel continues to be Intel, with a road map of standard microprocessors which assumes the computing world won’t change. This is how Intel missed mobile, a market it is only now making inroads in after a decade of failure.
There are some writers who insist the semiconductor company’s problem is management, but Andy Grove is not walking in that door, and the 20th century is in the distant past.
I believe the way for Intel to gain the customer focus it needs to succeed in today’s market is to separate itself from its fabs. Break the company up, one part designing chips and one part making them. Let the fab seek customers more actively, and schedule production based on its customer needs, rather than the roadmap. Tear up the road map, and build a new design house built around the marketplace. Let both sink or swim on their own.
The whole should be worth more than the sum of its parts. But it will take radical surgery to get Intel out of its doldrums. Until that happens only income investors, focused on a dividend yielding 3.1% at current prices, should be interested in the stock.
Cold Semiconductor Stocks: Qualcomm (QCOM)
In the market for the radio chips that power today’s mobile phones, Qualcomm is practically a monopolist. Other companies make radio chips, but Qualcomm controls the vital intellectual property needed to turn chips into services, and its royalties call the tune for the market.
This is a policy that governments naturally resent. Both China and South Korea have been fighting Qualcomm, and Qualcomm has reached favorable settlements. Now Apple Inc. (NASDAQ:AAPL) has gotten into the ring, and the outcome is not as easy to predict.
Supposedly Apple and Qualcomm are suing one another over patents and pricing, but the argument is really over the cost the mobile industry must bear for Qualcomm’s monopoly power. Apple has tried to bring the U.S. government in on its side, both sides have lawyered up, and the two sides have settled in for a siege.
As a result, since the start of 2017, Qualcomm shares are down 19%. The Apple suit shaved $10 billion off Qualcomm’s market cap, and resulted in it showing just $5 billion in sales for the quarter ending in March, against $5.6 billion a year earlier, as Apple cut off payments. Net income was also down, to $749 million from $1.2 billion a year earlier. The stock has recovered a bit from April lows because it’s assumed the feud will end.
The legal problems continued to dog financial results for the June quarter, with net income of $866 million, 58 cents per share, on revenue of $5.4 billion. The numbers beat analyst estimates but the shares fell another 5% to below $54 each.
Analyst Rob Enderle thinks Qualcomm has the upper hand in the legal war but that Apple could still “go nuclear,” turning its marketing department on Qualcomm’s brand.
Qualcomm’s royalty rates for its patents are based on the cost of finished goods, and since iPhones are pricey Apple pays than its competitors for the same rights. Apple also says it can’t play Qualcomm against other chip suppliers because of Qualcomm’s “no license, no chips” policy, which guarantees it will pay more that way. It’s an argument that has been repeated by the Federal Trade Commission, which filed its complaint against Qualcomm right before Apple filed suit.
Qualcomm is also seeking to close its $47 billion acquisition of NXP Semiconductors NV (NASDAQ:NXPI), its entry into the huge Internet of Things market, so the timing of the argument over its monopoly rights — patents, copyright and trademarks are all legal monopolies — is not good.
What investors need to remember, however, is that once it obtains peace, the dividends for Qualcomm will be substantial. This semiconductor stock may remain moribund for months, and to keep the price up Qualcomm recently hiked its dividend to 57 cents per share, a yield of almost 4%.
Peace will mean a higher price for the stock and, correspondingly, a lower yield, so if you’re looking for dividends now is the time to buy. If you’re looking for capital gains, wait for signs of peace, then buy with both hands. For now though, Qualcomm, like Intel, is strictly an income play.
Cold Semiconductor Stocks: Micron (MU)
These are the good days for Micron Technology, Inc. (NASDAQ:MU), but investors in semiconductor stocks need to be on their guard.
The Boise-based memory chip maker finds less competition than ever, at a time when demand seems to be skyrocketing. For the quarter ending in March it had net income of $894 million, 77 cents per share, on revenue of $4.6 billion. Earnings delivered June 29, however, fell short of expectations at $1.40 per share, when they were expected to come in at $1.50 per share, although revenue of $5.6 billion beat the estimate of $5.4 billion
So why did the company abruptly close its Lexar retail unit, which made things like memory cards and flash drives for the retail market?
The move shocked observers of the stock and will drop the number of competitors in the space to two, Sony Corp (ADR) (NYSE:SNE) and SanDisk Corporation (NASDAQ:SNDK). It took Micron most of July to return to its pre-earnings level of about $32 per share, despite a rising market.
Micron had named SanDisk co-founder Sanjay Mehrotra its next CEO in April. Mehrotra’s first big decision was to close Lexar. Maybe he’ll even move to sell the company — there are constant rumors of its being sold to Intel when business slows. But that will be after business slows. Those who buy today may not get back their investment, even on a buyout.
The problem for investors is that memory chips are a notorious boom-and-bust market. Micron has lately been on the boom end. The shares are up 37% in 2017. Micron was a hot stock during the last boom period in 2013-2014, too, exceeding current prices at $35 per share, but it crashed hard in 2015 and hit a 2016 low below $10 per share.
Since its founding with money from potato billionaire J.R. Simplot, Micron has experienced more ups and downs than a Mexican Telenovela, with multiple rounds of layoffs each time the computer upgrade cycle slowed, and a 2012 plane crash killing then-CEO Steve Appleton.
I have seen this movie before. Memory prices rise to meet each upgrade cycle, but production exceeds the market need and prices fall again, putting pressure on the balance sheet. Each new generation of memory costs more to produce than the previous one. Micron already has $11.7 billion of debt on $33.3 billion of assets.
The current boom for chip stocks will pass, and Micron will come under pressure again. If you do choose to buy, make like a potato and don’t take your eye off it.
Dana Blankenhorn is a financial and technology journalist. He is the author of the historical mystery romance The Reluctant Detective Travels in Time, available now at the Amazon Kindle store. Write him at email@example.com or follow him on Twitter at @danablankenhorn. As of this writing he owned no shares in companies described in this article.