NVIDIA Q3 2018 Pre-Earnings Analysis

This is not investment advice. The author has no position in any of the stocks mentioned. WCCF TECH INC has a disclosure and ethics policy.

NVIDIA is all set to report its quarterly earnings soon and as things stand right now, this is going to be one of the more interesting numbers reported in recent history. The result that NVIDIA (NASDAQ:NVDA) reveals tomorrow is highly significant and can be used to set the baseline risk of the company. Q3 2018 has mostly been without any significant crypto demand (not there was much for NVIDIA, to begin with) and also saw no major traction of its newly released Turing cards. So the question on everyone's mind is of course, can Jensen pull a rabbit out of his leather jacket and beat?

I will be going through some of the key points to keep in mind while dealing with a stock as ambitious as NVIDIA - the ones that could lead to a surprise beat and the ones that could lead to analyst expectations barely being met (or even an underperform). There are a lot of variables at play when talking about a quarterly result and NVIDIA (NASDAQ:NVDA) is no different, but the subtle nuances of the GPU market is something that analysts very rarely understand, which is why it is quite possible that the company would find it much more difficult to beat expectations this time than last time around.

NVIDIA A100 PCIe GPU Accelerator Gets Liquid-Cooled, Offers Same Performance While Consuming 30% Less Energy

Factors leading NVIDIA (NASDAQ:NVDA) to miss analyst expectations

  • Analysts are expecting disappointing earnings and chances have never been higher that they are right.
  • NVIDIA's newly released architecture, Turing, hasn't gained as much traction as older uArchs as Pascal did at launch. The obvious reason for this is that the price tag is something that has been very hard to stomach for most of the market.
  • The company introduced a brand new rendering technique (RTX) that are only supported by Turing cards and also added AI cores to speed up performance later down the line. The only problem? RayTracing and Tensor-based AI tech embedded in Turing cards isn't working yet. Only a handful of demos support it right now although some AAA devs have promised support for it in existing and future titles.
  • It goes without saying that the cost of R&D for Turing would have been very high. The company has essentially added 2 more engines (Ray Tracing and Tensor) to the gaming GPU, which previously consisted primarily of a Shader engine.
  • The lack of traction for Turing and high R&D is what makes this quarter very interesting. It would seem like all the odds are stacked against NVIDIA right now and how the company performs will establish a lower baseline of risk for investors. If Jensen can beat these odds once again, they can probably keep on doing so. If they can't, then we know which indicators to look out for going forward.
  • There is also a tangential issue related to the underlying process tech that they utilize from TSMC. Process tech has reached a sort of saturation point. 12NFF - the node Turing is based on - is an optimized version of the old 16nmFF node regardless of what the name may imply and the physics of process shrinks is getting harder (and more expensive). There is only so much performance you can squeeze out of new architectures if the physics of the die are limited. This is why companies like NVIDIA have to innovate new sources of performance. The Tensor cores in Turing are one such example but they have not been enabled yet.

Factors leading NVIDIA (NASDAQ:NVDA) to a beat in Q3 2018

  • There are various indications that Turing was deliberately priced the way it is and the company did not intend it to gain mainstream traction till Pascal inventory finishes. In other words, the lack of mainstream traction of Turing is exactly what the company intended and wanted right now. The niche market that can afford to pay the premium for the bleeding edge is tapped while everyone else helps exhaust Pascal inventory.
  • Pascal cards are flying off the shelves. If you are someone neutral who cannot afford Turing, you are not going to turn to AMD cards (which are not competitive at most price points), you are going to turn to Pascal cards. Any customer Turing loses, is in all likelihood, going to Pascal, its a zero sum game for NVIDIA either way.
  • The company has been acting in what is essentially a monopoly as far as gaming cards go. As I mentioned earlier, AMD is only competitive in a select few price points, for everything else, its NVIDIA all the way.
  • The company released new Turing based professional GPUs that feature Tensor cores. This is something that will allow the company to go head to head against Googles TPU, provide vastly superior value in its professional lineup and just help it take market share from competitors. Unlike the gaming segment, Turing here is actually well worth the price.
  • Jensen has a magical black leather jacket that helps him beat analyst expectations. (This is a legitimate reason and I take issue with anyone who thinks otherwise).

At the time of writing the company's stock is trading at $199, which seems to be a far cry from the $300 ambition the stock was chasing just a couple months back. Based on the factors above, we feel the stock is undervalued right now (ceteris paribus) and has high upwards potential with low downwards potential. That said, movements during earnings are usually volatile and can be unpredictable and a sell-off has historically happened during this time frame so certainty levels would be high only on medium-long term plays.

Disclaimer: I do not hold NVDA (or related/rival) stock and do not plan to do so for the near future. This analysis is provided "as is" for you to read and you assume all responsibility for any trading plays.