- Nvidia has sustained a record run, and has remained the best performing semiconductor stock in 2016.
- Upon exploring downside scenarios, I still walk away with the impression that the recent pattern of earnings beats will continue.
- However, the quarterly volatility will become more severe, as expectations are elevated and may eventually detach from reality.
After witnessing what has been a somewhat volatile Q3’17 earnings announcement , I felt it was necessary to update my stance on the NVIDIA (NSDQ:NVDA). After all, the stock is now up over 400% from when I had first initiated coverage on the stock back in 2014, and is up 268% from its 52-week low. I guess, one must wonder, just how long will the elevator ride last?
The company is posting up Stephen Curry like numbers every quarter. It’s magical days for Nvidia and CEO Jen-Hsun Huang, like the Steve Jobs era of Apple from 2008 to 2012. But, investors will need to keep a close eye on earnings in each consecutive quarter, as it will make or break returns depending on the outcome.
As Peter Lynch, would often say:
If you can follow only one bit of data, follow the earnings -- assuming the company in question has earnings.
I subscribe to the crusty notion that sooner or later earnings make or break an investment in equities. What the stock price does today, tomorrow, or next week is only a distraction.
Brushing up on some history
I’m not going to lie, it’s a little difficult task to paint a fair earnings/sales trajectory for Nvidia, as no one really thought the refresh to Pascal would translate into 54% y/y sales growth in Q3’17. But then again, the Maxwell family was the 4th generation 28nm series, and as such, Nvidia had stretched the lifespan of the 28nm node for four consecutive years by tinkering with clock speeds, CUDA cores, GPU design, driver support, etc. Of course, there was only so much headroom to these various approaches, the performance/watt only marginally improved, and with many mid-level GPUs capable of running 1080p AAA video game titles, the GPU market stagnated from 2012 to 2015.
We then roll into 2015 and 2016, where 2K and 4K gaming monitors started to become more prevalent, which increased the pixel density by 2x to 4x when compared to 1080p, which required 2x to 4x the graphical computing power to render the added pixels at stable frame rates. However, for the first time in quite a while, we witnessed high-end Nvidia GPUs ranging from the 1080TI to the Titan X incapable of running games at fully maxed out settings and resolutions.
Typically, game developers design a game to run efficiently across mid-end graphics cards with some room to throttle graphics settings higher, but only to a point where high-end cards can run at stable frames. However, the PC ecosystem shifted to higher-end configurations with 4K and VR driving performance needs much higher than in years past.
The game developers weren’t setting the usual pace of innovation, it was the broader ecosystem that was driving PC refresh, which is why Nvidia experienced more of a transitory benefit from other technologies in conjunction with more graphically intensive titles. The computing needs being not fulfilled by Maxwell generation cards was what caused the massive shortages for GTX 1080 and 1070 series cards upon release.
So, what’s in store for the future?
Pascal doesn’t run the most demanding games at stable enough frames to label them true 4K-capable graphics cards, so you’d have to run them in an SLI configuration, but that drives the price to $1,200 for a pair of high-end cards. The sweet point for the enthusiast segment is somewhere around $400 to $600, so the pricing discrepancy will force the enthusiast segment to wait for another year or so before performance/pricing catches up with 4k/ultra-graphics settings. Only after this happens the market for enthusiast gaming will rapidly expand, because it’s now more expensive to run the latest game titles with fully maxed out settings.
Driving ASPs higher will be difficult, and it will require volume purchases in the mid/high-end at roughly $400 to $600 for revenue to sustain Nvidia's current pace of growth. However, upon fully saturating the market with cards that are reasonably affordable and can run 4K at maxed out settings, the demand for high-end cards will likely decelerate, as demand for high-end PC parts is driven by the most demanding use scenarios. And once those use scenarios are fulfilled at reasonable pricing tiers, we run into a situation where enthusiasts upgrade less frequently.
Sort of like… what happened in the CPU-space where Intel (NSDQ:INTC) would launch a new line-up of CPUs with a muted response, as Haswell-series CPUs were sufficiently capable even in the most demanding use-cases. So, no one would abandon their Haswell series CPUs for Broadwell, Skylake, or Kaby Lake for that matter. CPU refresh slowed to a four to six-year cadence, which is why Intel’s PC segment has been stubbornly stagnant. There’s not enough innovation to drive purchases of faster CPUs, because no one can justify the performance improvement, for what it costs, and there aren’t enough useful applications for those faster CPUs anyway.
It's like replacing a high-end car every year. Just because BMW releases the same M5 Series with a modest improvement in speed, the relatively affluent won’t go and replace their car every year. The performance gains in automobiles, or CPUs for that matter have become very marginal, and without the added usefulness of faster freeway lanes, or in the case of CPUs – more demanding applications, there’s no reason to replace either of them.
Likewise, Nvidia is caught in a scenario where it can keep replacing its graphics line-up to meet the most cutting edge applications, but once performance and pricing starts to catch up to those applications, the enthusiast segment will throttle back their spending, and we’ll be stuck in a more stagnant GPU environment like 2013 or 2014 for that matter.
So, what does this mean for investors?
We haven’t witnessed the GPU-market stagnate the way CPUs have, but given enough time, the cyclical nature of hardware will eventually rear its ugly head. While the potential for cyclical weakness does loom over the heads of investors, I also believe investors are pinning their hopes on emerging categories like datacenters and autonomous vehicles where silicon content for GPUs are rapidly expanding. So, the TAM for GPUs could expand at a much quicker pace than historical trends would suggest, which is why investors are piling onto the Nvidia momentum trade.
But then again, the stock will always pause on guidance misses. Buy-side analysts stuttered upon the announcement of Q3’17 earnings, as Nvidia guided $2.10 billion sales for Q4’17, which compared to $2.004 billion in Q3’17. Therefore, the management outlook implied that the one-off impact from the launch of Pascal-generation cards is cyclical, and q/q growth to the tune of 10-20% may not be as sustainable.
Notwithstanding the flat guidance, investors are chasing the stock higher, even into the $90s. That’s just crazy… but then again, it has the semblance of another Netflix, Amazon, or Facebook story. So, it’s still racing off to the moon.
While, playing devil’s advocate isn’t what most readers are looking for. I realize… and most will agree that we’re stuck in a super-cycle. But, this super-cycle won’t last forever, and it’s always worth noting what could potentially go wrong in this momentum themed play.
Given enough time, momentum themed investments eventually slow and will correct themselves upon downward revisions in various buy/sell-side financial models. So, while I keep thumping the Nvidia story, there will also come a point where I too will change my stance upon saturation and declining y/y comps.
That time has yet to come, and I doubt it will be next month or next quarter. Heck, not even next year. However, it’s worth keeping a close eye on the stock, because there will come a point where the bag blows out from underneath the stock upon a series of crushing earnings announcements. Therefore, upon signs of weakening fundamentals, investors will need to liquidate their positions to protect loss of capital. The situation is fairly tenuous because expectations are just that high. I continue to reiterate my buy recommendation on Nvidia.
Looking to invest in technology companies? Here are our latest Top Tech Stock Picks which have outperformed the market consistently.