This article was originally published on September 28, 2017. It has not been updated or revised.
Charles Mizrahi has mostly pitched stocks that have a “value” bent in the past, but his latest ad for the Park Avenue Investment Club is much more “growthy” in nature. Here’s how it opens up:
“World’s First Trillionaires Will Come From THIS
“Google just started using it, and it made them an extra $9 BILLION in revenue…
“It’s expectd [sic] to be 18 times bigger than smartphones, tablets and PCs… COMBINED
“And it will unleash a $50 trillion new wave of wealth…”
The pitch is one of many that have quoted Mark Cuban’s sentiment that the first trillionaire will be an artificial intelligence entrepreneur, somehow harnessing the Internet of Things and Big Data and AI to create the next goliath… and that sentiment is widespread in the investing community, most experts I’ve heard talk about this echo that sentiment that deep learning and AI will be the basis for the next dominant silicon valley company (though they all tend to note that yes, Alphabet and Facebook and others who are investing heavily in AI will also benefit from this next wave of technology).
Artificial Intelligence and Deep Learning can mean many things, but the basic investment premise is that computers that can “learn” from new data that they collect, without having to be reprogrammed or updated by a person or a new reprogrammed algorithm each time something changes. That gives the capability for computers to become exponentially more capable, and enables not only self-driving cars and image search within videos, which are some of the current core applications of AI (a car gets better at spotting bicyclists after the computer has seen ten million bicycles on the road that look and behave slightly differently, for example, not just because a programmer tells it what a bicycle looks like), but also great new and unforeseen capabilities in the future… particularly as the vast network of Internet of Things devices creates more and more data that can be processed by the next generation of deep learning machines.
Mizrahi also takes a moment, as pretty much all newsletter pitchmen do, to boast about a few of his greatest performers — including Huntington Ingalls Industries (HII), which he did indeed feature in a teaser ad back in the fall of 2012 (it’s up 400% or so since then, and he says he still expects great things from that Naval contractor — that’s one of my “shoulda” companies that was appealing at the time but that I didn’t buy). So yes, like pretty much anyone in this business (or any investor, frankly) he has had some big winners.
I don’t know what his overall track record might be, but I do also own the stock he has pitched most aggressively over the past several years, Fairfax Financial (FFH.TO FRFHF), and I still like it… despite the fact that it has trailed the market pretty dramatically over the past year (and has been a mediocre stock for an extended period, returning roughly as much as the S&P 500 since the end of the financial crisis).
So that’s the big picture and a little background. What’s the actual investment that Mizrahi pitches as a play on this?
It’s some kind of chipmaker, and he compares this to investing in Intel before the personal computer revolution took hold (and to other successful semiconductor stocks, like NVIDIA)… but let’s dig some clues out of the ad so we can see what he’s talking about.
Here’s a taste:
“The Single Best Opportunity of the Next Decade
“Deep learning is the future of ALL technology.
“The time to get in is NOW.
“And at the nexus of it all is a tiny component made by a cutting-edge company that could hand early investors 10, 15, even 30 times their money.”
And, in case that isn’t tantalizing enough…
“… a remarkable — yet, overlooked — company is at the forefront of it all.
“If you were to make only ONE technology investment in the next decade, this company should be it.
“I have never been more certain about anything in my entire life.”
Every time I see language like that “I have never been more certain” bit, I have to remind myself that this is probably the ad copywriter talking, not the investment analyst. I’m sure Mizrahi isn’t putting all his assets into this one stock, which is what you would logically do if you were more certain about it than you had ever been about anything else in your entire life (choosing your spouse, loving your children, serving your country, whatever). Perhaps he has high conviction in this stock, even more conviction than he usually does about his investment ideas, I don’t know… but I find certainty to be elusive when it comes to investing, and probably more likely to lead to hubris and disappointment than to riches.
But anyway, what else do we get by way of clues regarding this “first trillionaire” stock?
First, a bit more of the background that Mizrahi trots out to explain his understanding of AI…
“… this breakthrough technology is already powering advancements in a whole host of industries. It’s being used to:
“Identify potential treatments for Ebola, multiple sclerosis, and various types of cancer.
“Recognize potentially threatening climate events.
“Help farmers detect crop disease BEFORE it spreads.
“And, heck, even predicting trends in the financial markets.
“And we’re just starting to scratch the surface of the potential applications…
“Experts say that there will be roughly 50 BILLION devices connected to the internet by 2020.
“And I’m not just talking about your smartphones, watches, or tablets, but also internet-enabled thermostats, light bulbs, fridges, baby monitors, toasters, windows, and more…
“The term for all these connected devices is called the Internet of Things, or IoT….
“Without deep learning, the $7 trillion Internet of Things can’t reach its full potential.
“And I’m not just talking about Netflix giving you better movie recommendations or about Pandora giving you more songs that you like…
“By improving the speed and accuracy of data analysis, deep learning can optimize productivity (and profits) across nearly EVERY industry on the planet.”
He does note the vast amount of investment going into this sector… which, I’d note, should serve both to intrigue us, as Mizrahi intends, and to make us wary of the many, many other companies that are trying to fight for leadership of this market:
“CB Insights, a venture capital intelligence firm, reports that more than $18.4 BILLION has been invested into AI-related startups over the last five years.
“And investment continues to accelerate at a staggering pace.
“Here’s what it boils down to…
“The richest, smartest investors in America are loading up, and I suggest that you do the same — immediately.”
OK, so we obviously can’t buy most of those venture companies that are getting that $18 billion in funding… and most of them will probably fail, anyway… but what is it that he thinks we should buy?
“I’ve found a much smaller company that is helping power Amazon’s AI unit.
“And get this: It also happens to be Amazon’s fastest-growing business segment!
“Sales for this firm have been growing for the past seven consecutive quarters with no slowdown in sight.”
OK, so it’s got sales growth and it’s working with Amazon. What else?
“… computer processors were NOT designed for deep learning. And it’s shaking the entire microprocessor industry to its core.
“These new chips facilitate the growing wave of machine-to-machine communication for things like connected cars and smart appliances.
“Wired is calling it ‘a new breed of chip that could accelerate AI from inside smartphones and other devices.’
“And MarketWatch says that these chips ‘are the pickaxes and shovels in a potential robotic-car gold rush.’
“These supercharged chips will speed up deep learning in everything from tiny devices to massive data centers.
“So, to get in on the action, you will want to own stock in a company that makes these chips.”
Simple enough logic there, though we should probably ping our brains here for a moment to remind ourselves that there’s almost never “just one company” that makes the chips for a particular application. Newsletter copywriters excel in drawing that connection from a powerful trend to a single company, and usually instead of that one bold line we’re probably really looking at a giant web of connections between hundreds of companies and suppliers… particularly with a big trend like “deep learning” and AI.
So with that caveat, what other clues do we get from Mizrahi’s ad?
“Formed in Silicon Valley in the mid-’80s, this company’s CEO is a 35-year tech pioneer, and he’s using his industry contacts to lock in some very impressive contracts.
“Investing in this Silicon Valley pioneer today will perfectly position you to enjoy a ride on the greatest wave of innovation that history has ever seen.
“In the past year, internet giants Amazon and Baidu have signed on to use this company’s unique chips. In fact, these chips are what’s powering Amazon’s fastest-growing business unit.
“Electronic Design says that this company’s chip is a ‘jack-of-all-trades [and that] engineers can reprogram them to handle tricky computing jobs like speech and image recognition better than general-purpose computers sold by the likes of Intel.’
“In fact, this company is THE market leader when it comes to making the chips. And this also makes them one of the last great buyout targets in this explosive realm.”
Enough? Indeed. I had to move the leaf blower out of the way to get it cranked up on this chilly New England morning, but the Thinkolator did indeed fire up right away for me — and after shoveling in those clues, the answer came out nice and quick: This is Xilinx (XLNX).
Xilinx was indeed founded in the mid-80s (founded 1984, went public in 1989), and the company was created to commercialize the first field-progammable gate array (FPGA) chips — which are, essentially, chips that can be programmed to perform different tasks after they’re manufactured (instead of being manufactured for one particular set of processing tasks). FPGA chips are indeed being used to boost the artificial intelligence capabilities of data centers and in other applications, and Apple even surprisingly included one in the iPhone 7 last year that was widely seen as a possible “be ready for AI or virtual reality” placeholder.
The challenge, of course, is that Xilinx does not own the idea of the FPGA — they were the first to commercialize this concept and sell FPGA chips, and they are still the industry leader according to everything I’ve read, with roughly 50% market share, but lots of other semiconductor companies sell FPGA and other programmable semiconductors now. The second place player in this market is the particular concern, because that’s Altera, which was bought by Intel a couple years ago and thus enjoys Intel’s marketing and manufacturing might (Xilinx is a fabless semiconductor company, meaning they contract out to others to build the chips… in fact, they’re given credit for inventing “fabless” as a business model, which gives them great margins but somewhat less control). The other meaningful players are Microsemi (MSCC), QuickLogic (QUIK) and Lattice (LSCC) — Lattice has been in the news recently because the US government blocked their acquisition by a Chinese state-funded investment firm, and the other two have very teensy FPGA businesses compared to Xilinx and Intel/Altera.
The growth is not particularly sexy, however, and though the trend has been toward growth their annual revenue has been in the same general neighborhood ($2-2.5 billion) for more than five years now. The financials actually paint an interesting picture, because this is a company that has not increased sales or improved margins or grown earnings meaningfully in at least three years… yet they’ve seen their PE ratio and their Price/Sales ratio grow dramatically.
Or to put it more bluntly, the company hasn’t gotten any better over the past three years, it has just gotten a higher valuation… at least, that’s how it looks in the raw numbers. That doesn’t incorporate the story at all, and it doesn’t incorporate this notion that FPGA’s and other advanced programmable chips will become more important as AI takes its place in more products (and in more data center applications), so if you want to like Xilinx, that’s where you have to find the reason: Are they going to continue to be a dominant provider of FPGA chips, and will that market grow dramatically over the next few years?
I can’t answer that, but I can say that the analyst forecasts are pretty underwhelming for a company that trades at Xilinx’s valuation. Xilinx at $70 is trading at at about 28X forward (FY2018) earnings, and those 2018 estimates represent 9% earnings growth over last year (XLNX is in Q2 of its fiscal 2018 right now). Further estimates are for 9% earnings growth in 2019 and 1% earnings growth in 2020 (though the 2020 estimates are from only three analysts, so perhaps they’re just the most cautious ones — the other years average about 15 analyst forecasts).
That’s not necessarily bad, and the analysts could be underestimating the potential for some kind of explosion in FPGA chip demand, but my initial reaction is that it’s not good enough for a company that’s trading at 28X earnings — that means you’re buying at a PEG ratio of about 3 (meaning, the PE ratio is about three times the expected earnings growth rate). That doesn’t usually work well unless the “G” turns out to be a lot better than expected, or unless it’s a really stable slow-growth industry where folks are happy to buy for the dividend and ignore the growth rate. Xilinx just started to pay a dividend last year, the yield is currently about 2%… and they do have the capacity to grow the dividend, it currently represents between 50-60% of earnings and they have a couple billion dollars in cash on the balance sheet so they’re certainly in no distress.
Having a high PEG ratio doesn’t mean the stock is guaranteed to fail, mostly because forward PE ratios can certainly be wrong if analysts are too cautious or are just missing something, or because investors sometimes get really excited about “future” stories whose earnings haven’t caught up to the stock yet… but it means the risks are pretty high because you’re paying a steep price relative to the consensus expectation about future growth rates.
I would say that the biggest reason for optimism is that Xilinx has been boring and up-and-down with its business for so many years, and FPGA has been predicted as the “next big thing” for so long without becoming a real mainstream story, that perhaps this means the analysts have gotten cynical and will miss the next big growth spurt — analysts don’t often get very far ahead of the news, and don’t typically want to overshoot on their earnings forecasts, so if there’s a big turn coming for Xilinx, as Mizrahi seems to expect, perhaps the analysts just don’t believe it yet.
So by all means, go forth and research Xilinx — see if you can find a reason why the earnings will grow much faster than analysts anticipate, or if you think there’s a likelihood of a premium buyout offer. Those are the two things that would make it reasonable to pay $70 for the stock, in my opinion, and I don’t see them in my 20-minute tour through their numbers… but that doesn’t mean they aren’t there.
If you’re looking for optimism about FPGA’s, and the kind of chips that Xilinx makes, start out with this Wired story from last year or this story about Amazon’s XLNX partnership and the possibilities it might bring… or check out the recent piece about Google’s new Tensor Processing chips here (Google’s new chips, which have sometimes been talked up as “data center killers,” aren’t FPGA’s, but in part address the same issues caused by increasing demands on AI processing from data centers).
Notwithstanding the interesting possibilities of those stories, I don’t know enough to be confident in them yet. Xilinx is a good company, with a strong business, good leadership, a long history of dominance in their sector, and a decent 2% dividend… but my initial sentiment is that it’s too expensive. What say you? Throw your wisdom on us with a comment below.