All-In on Intel
In an article for Barron’s, BMO Capital Markets analyst Ambrish Srivastava recently was quoted as saying “I think investor perception is that the whole market is going A.I. and Nvidia is the only way to play it.” His words while not prophetic, do reflect current market sentiment that Nvidia is on the forefront of A.I. and other companies have to play catch-up within the space. This would also explain why Nvidia is the best performing stock in the Standard & Poor’s 500 index, up 200% on the year. However, while Mr. Srivastava was making the point that the market perceives Nvidia as the only way to play the A.I. industry, his own personal views are that Intel, while later to the game as far as market capitalization, is actually better positioned to take the reigns as the preeminent leader in A.I. technologies within the next several years.
Nvidia was first to the A.I. party because of it’s longstanding pedigree in gaming. Anybody who is or was a computer nerd (such as myself) could attest to the fact that if your gaming rig did not have an Nvidia chip, you had no clue what you were doing. Basically as a gamer, your Nvidia chip would react to changes more seamlessly, thus giving you the advantage on your opponents. Video Games are seen as a precursor to a lot of A.I. concepts, and thus have given Nvidia the leg up on other chip makers when true A.I. went from a futuristic idea to an actuality these past two years. However, we are only at the beginning of the A.I. era, and the kings of A.I. have a long way to go before they get crowned.
Within the A.I. umbrella, there are three subsidiaries that every tech giant in Silicon Valley is looking to innovate in. These three areas of A.I. are autonomous driving (vehicles that drive themselves) machine learning, and deep learning. Machine learning is the process in which machines take technical data and algorithms to make predictions of real-world events. Deep learning is even more impressive, it uses ‘neural networks’ which are inspired by how organic brains function, and uses trial-and-error processes to learn and perfect new ways of accomplishing goals. Deep learning can be thought of as the A.I. equivalent of a Juilliard Pianist who spends 14 hours a day perfecting every aspect of Rachmaninoff’s Piano Concerto no.2 through uninhibited repetition. The only difference is the A.I. will know more about every aspect of the Concerto in a couple months than a master pianist will know in his/her entire lifetime of playing the piece.
So how will Intel make a significant impact, more than even Nvidia, in the three major aspects of Artificial Intelligence (Autonomous driving, machine learning, and deep learning)? It already has begun its ascension in the ranks of all these technologies through systematic and targeted acquisitions and partnerships. First, Intel Plans to release a new line of A.I. chips developed in collaboration with Facebook later this quarter. Intel also purchased Mobileye, who is the preeminent firm in autonomous driving, having a miraculous 70% of Market Share in the space. In addition to its acquisition in Mobileye, Intel also acquired Movidius last year, which makes vision chips for drones as well as acquiring Altera last year, who lead in ‘field programmable gate arrays’ (FPGAs). All you really have to know about FPGAs is that they can be used alongside Intel chips to speed up processing in servers. As if this wasn’t all enough to make Intel a behemoth in A.I., they also purchased Nervana which is a Deep Learning start-up with a lot of promise. This quarter using the aforementioned acquisitions, Intel plans to launch the Nervana Neural Network Processor family, bringing them more on the cutting edge of deep learning and other A.I. applications.
The Nervana Neural Network family of processors will be involved in a new Microsoft A.I. project code-named Brainwave as well as 14 autonomous driving contracts as of today. Because there are so many applications for this new era of chips, expect multiple players to thrive in this area in the upcoming years. Meaning the market isn’t mutually exclusive- both Intel and Nvidia can make significant gains on their respective stock prices in the upcoming calendar year. However, I do believe that Intel at this junction has a higher upside then Nvidia. I think a lot of Nvidia’s potential in the A.I. market is already built into the stock price, while Intel’s potential not only isn’t yet fully realized, but has a higher plateau in general than Nvidia with all of its recent acquisitions and partnerships.
Of course, there is also always the question of cash. According to consensus estimates referenced by Barron’s, Intel could generate $14 billion in yearly free cash, as opposed to an estimate of just $3 billion for Nvidia. In estimates also referenced by Barron’s, research-and-development spending for Intel will surpass $10 billion this year, versus approximately $2 billion for Nvidia.
And now it’s time for the story within the story- Intel has another partnership with chipmaker AMD to develop new AI chips in a stark opposition to Nvidia’s new-found dominance. Jim Cramer described it best by facetiously saying Intel just kept AMD around instead of running it out of business to keep anti-trust committees off its back. Truly AMD, aside for maybe a single quarter back in 2007, was never actually competition to Intel. I never liked the stock prior to this, but AMD may be a nice pick-up and dump considering its recent partnership with Intel. Of course, I already mentioned how Intel and Nvidia are studs, and AMD has always been the red-headed step son of the group. It’s partnership with Intel basically gives it certain upside.