Wolf's Stock Pick for February 8, 2016 - Nvidia (NVDA)

I used to think of myself as a Dividend Investor (see last week's post for some basics about dividend investing).  I still think that dividend investing is a good, safe way to make money - and I believe that once you have enough money in your investment accounts, you should convert every stock you own to a dividend paying stock so that you can retire and live on the dividend income alone!  We'll talk more about that particular retirement strategy in a later blog - meanwhile, today, I'm going to talk about a speculative stock (one that you buy hoping to make money when the stock increases in price) - NVIDIA (NVDA).  While NVIDIA does in fact pay a dividend (roughly 1.2% at today's price per share of about $26),  I don't think of it as a dividend stock per se - but I do think you should buy some, and here's why.

You've probably already know of NVIDIA as a  manufacturer of graphics cards for computers.  They build these cards using their own proprietary GPUs (graphical processing units).  They also manufacture a System On Chip (SOC) that powers the NVIDIA-branded Shield Android tablet (which is more than just an Android tablet - it's a gaming system in disguise):

Nvidia Shield.  From https://shield.nvidia.com/

The core product of NVIDIA is the GPU, or Graphics Processing Unit), and they sell it in several ways - graphics cards, SOCs, etc.  And a GPU can be used for more than just generating graphics - GPUs are actually more powerful than CPUs (the main chip driving most computers) for certain types of processing - specifically, batch operations on large data sets, such as linear algebra computations and statistics.  This is because a GPU has, rather than 2 or 4 cores, as many as hundreds of cores, all performing the same function in parallel.  GPUs are being used to create supercomputers that in turn can run neural networks (machine learning) programs, bioinformatics analysis, computer vision and image analysis, and even oil exploration algorithms.  GPUs were actually the fastest way to mine bitcoins until they were replaced with even more specialized field-programmable gated arrays (FPGAs).

Okay, you're saying, I get it.  GPUs are cool.  But why should I buy NVIDIA stock today?

Great question - I'm glad you're following along!  The answer is that GPUs, for the reasons mentioned above, are going to start showing up in places other than your home computer, gaming console, or university supercomputer, and it's primarily going to be in two of the applications I mentioned above:


Machine Learning

Machine learning is used to make "smart" computer software called a neural network - we're not talking AI, or artificial intelligence (not yet, anyway), but machine learning has already been used to help doctors diagnose illnesses, choose which medicines to treat an individual with based on their own genetic information, and predict things like the outcome of presidential elections or stock market behavior.  In more bizarre examples, neural networks have been used to simulate the brain of the nematode C. elegans, and to create an autonomous quadcopter drone  using a simulated bumble bee brain as the "pilot".  Closer to home, you probably interact with machine learning software every day - most likely on your smartphone.  That's right - Siri, Cortana, Google Now, and Amazon Echo are all powered by machine learning.  Today most of the actual processing happens on a remote server, which is why you can't use Siri or any of the others without being online - but that will change as we find more uses for these intelligent agents in our homes.  For example, Apple's (AAPL) latest incarnation of the Apple TV uses Siri to let you talk your way through the TV listings, searching for Sci Fi movies starring Rutger Hauer, for example.  You can buy things directly from Amazon by telling your Amazon Echo what you need.  You can even voice control your smart lights and your thermostat, if you have the right hardware.  And as more of our homes become internet-enabled with such Internet of Things (IoT) gadgets, it's very likely that you'll be able to control them all through Siri or Cortana.  And as we add functionality to the neural networks running everything, they'll need more and more processing power - and that means GPUs.

But that's not why I think you should buy NVIDIA - while there will be increased use of GPUs for machine learning, I don't think that alone is what's going to make NVIDIA pop.  Let's add one more component to the mix:


Computer Vision and Image Analysis

Computer vision is defined by Wikipedia as "... acquiring, processing, analyzing, and understanding images and, in general, high-dimensional data from the real world in order to produce numerical or symbolic information..." (https://en.wikipedia.org/wiki/Computer_vision).  In simpler terms, computer vision is the application of cameras to capture still (or moving) images and passing them to sophisticated algorithms that can process those images (or videos) to extract information.  In the lab where I work, we have used computer vision to capture the limb motion of infants, analyzing the motion patterns to detect anomalous neuromuscular syndromes.  Mac OS X uses image analysis to detect faces from your pictures in the Photos app.  Google's Deep Dream combines computer vision and neural networks to process images and generate new images that are eerily reminiscent of something that Salvador Dali or Edvard Munch - or possibly H.R. Giger - would come up with.


The computer isn't "thinking", of course (and what encompasses thinking is still not clear - consciousness might be an extension of modeling behavior, and some think that image processing actually is a long step towards enabling true non-biological thought), so it doesn't do anything other process the image following the algorithm it's fed by the programmers.  Google Dream is at the extreme end of  computer vision, employing a true neural network to process image.  Nevertheless, the ability to teach a computer to capture images and videos and extra information from them is powerful - and it's this second piece of the puzzle that makes me think NVIDIA is a must-have.

What do you get when you combine neural networks with computer vision?  It's (probably) not the killer robots or singularity event that science fiction has prepared us for.  While the singularity probability is coming, it's not why you should buy NVIDIA.

So why have I led you so far down this garden path?  Why am I recommending you buy NVIDIA?

The answer:  autonomous cars.

Sigh.  Again you laugh.  And yet, it's true.  Autonomous (self-driving) cars are coming.  Tesla (TSLA) Motors is probably the most obvious example, but most other major car manufacturers - as well as other companies that aren't specifically in the automotive industry - have leapt on the bandwagon, providing cars that help you parallel park, manage freeway driving and lane changing, brake for you proactively when they sense you getting too close to another vehicle - the examples are numerous.  In fact the automotive industry itself is due for a shakeup once we get fully autonomous vehicles - imagine a future where nobody has to own a car anymore, because Uber and similar services make it possible for them to get a car anywhere, anytime - and more importantly, just as much car as they need (a convertible for the weekend, or a coupe for the daily commute) at the push of a smartphone app button.  Think of the money you'll save by not having to own and maintain a vehicle that spends most of each day just sitting....but I digress.

Google's self-driving car.  From https://upload.wikimedia.org/wikipedia/commons/1/14/Google_self_driving_car_at_the_Googleplex.jpg

NVIDIA's GPU technology can manage both the neural network processing and the computer vision required to make autonomous vehicles a reality.  And while there certainly are other companies manufacturing GPUs - notably AMD (AMD), Intel (INTC), and, remarkably enough, Apple (AAPL) -  NVIDIA has one huge advantage: they're already established a beachhead in the automotive industry.  Manufacturers like Tesla (TSLA), Audi (NSU), GM (GM), Honda (HMC), and Hyundai (KRX) are, or will be, using NVIDIA GPUs to drive their in-dash entertainment systems.  And the NVIDIA Drive-PX2 series of chip, described as a dedicated autonomous vehicle AI platform, and ostensibly one of the most advanced GPU systems available - has already been selected by Volvo (who recently claimed that within just a few years there would be no more driving-related deaths in their automobiles.  A bold assertion, but given Volvo's safety record, I'm willing to bet that they really think they can make this happen - and it's all going to be the result of cars that don't have to wait for a driver to make decisions.

Most of the manufacturers working on autonomous vehicles are being very close-mouthed as to the actual hardware/software being used to drive their vehicles, but it's a safe bet that NVIDIA is involved.  In fact, it's so "obvious" that NVIDIA is going to be a huge player in this market space in the next decade (Elon Musk himself said that fully autonomous cars are only two years away) that many savvy investors have already started snapping up NVIDIA shares - if you look at the 5 year stock chart for NVIDIA you can see a nice general upwards trend.

That being said, it's not too late to get in on the action - there's plenty of upside in this industry, given that we don't even have the first fully autonomous car - yet.  And as always, you can't really predict the future, only guess as possibilities - but smart money seems to be betting that when it comes to self-driving cars, NVIDIA is going to be along for the ride.

---------------------------------------------------------------------------------------------------------

Disclaimer - I am not a financial expert and I am not responsible for any losses - or gains - you may make if you make decisions based on the information posted here.  If you do make money, please feel free to let me know!

Disclosure - Wolfgang Rumpf owns shares in Apple (APPL) and NVIDIA (NVDA) mentioned above, and plans to purchase shares in Tesla Motors (TSLA) within the next 12 months.