To say Nvidia Corporation (NASDAQ:NVDA) did well last quarter would be a considerable understatement. Nvidia shocked the world Tuesday evening when it unveiled its Q1 results. Earnings of 85 cents per share handily topped estimates of only 67 cents per share of NVDA stock, while revenue of $1.94 billion topped analysts’ outlooks of $1.91 billion. Moreover, per-share income was up 85% year-over-year on 48% growth in sales.
The stock responded accordingly. All told, Nvidia stock advanced nearly 18% the next day, with pleasantly surprised investors racing to get into a name they’d only been lukewarm on since the beginning of the year.
The big-time growth and earnings beat does lead to one overarching questions though … how did this tech company with respectable but seemingly limited prospects manage to do so well?
The close-second follow-up question: Can it keep doing it?
An Unappreciated Technology Comes Into Its Own
Those who know a little about Nvidia will know its core business has been graphics processing hardware — the piece of a computer that optimizes what’s displayed on the screen. Most casual tablet and computer users don’t give that aspect of the underlying technology a second thought. Hardcore video gamers do though, since most games require smooth, fast image rendering to make the game playable and glitch-free.
Thus was born the graphical processing unit, or GPU … a term/acronym Nvidia coined in 2000.
While GPUs have been getting better over the course of the past couple of decades, the potential use of their underlying technology was never fully appreciated until very recently. As it turns out though, the ‘brain’ of the GPU is very fast, making ideal use of an idea called floating point operations.
Floating point operations are simply higher-level math functions. A GPU processor handles decimal-based calculations much faster than a central processing unit, or CPU can. CPUs, though they’ve come a long way as well, were developed from the ground up to primarily process whole integers. That is to say, CPUs can handle a calculation like “2×3” much better than it can handle a calculation like “2.47×3.16.” GPUs, however, can handle “2.47×3.16” like a champ. [That’s the very Q&D explanation, unsatisfactory to most computer scientists, but more than adequate for the average investor.]
At the same time, the concept of parallel processing has caught on, and become something of a must-have for heavier-duty computing.
At the onset of the computer industry, while a processor could perform functions very fast, it could only do one computation at a time. About a decade ago, the concept of computer cores was launched, effectively turning one CPU into two, four or more processors and vastly improving their speed. This is parallel computing, and it has opened a lot of doors.
While the commonly accepted means of several simultaneous calculations had mostly focused on splitting one CPU into multiple cores, the latest evolution of the idea is just lining up multiple physical processors and making them work as if they were one super-powered unit.
Now, for those who know the Nvidia story rather well and recognize it has made a huge splash in the field of deep-learning artificial intelligence, all of a sudden things make a lot of sense. Nvidia’s GPU hardware is more than capable of handling the higher-level number crunching that true AI demands, as it makes full use of parallel, floating point computing using its GPU technology.
Looking Ahead for NVDA Stock
As for the second question NVDA stock owners are asking (Can Nvidia keep up this growth pace?), the answer is yes, it can.
CUDA is an acronym for “Compute Unified Device Architecture” — a language invented by Nvidia to use GPU hardware in a variety of applications. Deep learning is one of those uses. A variety of enterprises are using the CUDA platform, including Microsoft Corporation (NASDAQ:MSFT), Alphabet Inc (NASDAQ:GOOGL, NASDAQ:GOOG) and Amazon.com, Inc. (NASDAQ:AMZN). They’re using Nvidia’s tech because if they want to make artificial intelligence a marketable, functional reality, there aren’t any other viable options as well supported.
Perhaps more important to current and would-be owners of Nvidia stock, analysts are (finally) starting to take notice of just how far CUDA and the floating point operations it makes possible with just a few strokes on a keyboard have come.
It matters. As Jefferies analyst Mark Lipacis recently commented:
“This supports our thesis that the center of gravity of computing is drifting toward a parallel model from a serial one. We think the Street is underestimating the secular nature of this shift and how it will benefit Nvidia.”
It’s something else Lipacis wrote, however, that could really put a well-deserved spotlight on NVDA stock. He said “Nvidia + CUDA appears to be emerging as the ‘Wintel of AI,'” referencing how Microsoft Windows and CPUs from Intel Corporation (NASDAQ:INTC) symbiotically revolutionized the computer industry in the 90’s. He added that Nvidia’s CUDA platform is “years ahead” of its competition.
There’s nothing in the market’s ether to suggest Lipacis’ assessment is wrong.
That’s not to say there aren’t valuation concerns. NVDA stock is priced at a fairly frothy trailing P/E of 49.8, and most analysts noticed that Nvidia’s expenses were up more than expected last quarter. It doesn’t really matter though. The AI era is still in its infancy, and Nvidia is well positioned to dominate it. Investors are generally willing to pay a steep premium for an undisputed champ.
As of this writing, James Brumley did not hold a position in any of the aforementioned securities.