NVIDIA Stock Crashes - It's a mystery to everyone

  • Registration is closed without referral. This is a website about Internet drama.

Pickle Pirate

Pickles are just undead cucumbers.
kiwifarms.net
Joined
Dec 19, 2014
https://wccftech.com/nvidia-q3-2018-results-detailed-breakdown-heres-the-why-of-the-crash/

NVIDIA stock collapsed in the after-market hours by more than 17.5% and was trading at $166.9 pre-market at the time of writing. This is a 180-degree inversion of the $300 ambitions the stock was chasing just a couple of months ago. The bears are in charge of the stock for now, but are they here to stay? We will look at what went wrong with NVIDIA’s disastrous third-quarter earnings after a streak of record-breaking quarters and what the rational outlook to the stock should be.

NVIDIA-Corporation-NVDA-NASDAQ.png


This earnings analysis will be presented in a slightly different format from before, I will be focusing on fundamentals and dissecting the numbers this time, with very little focus on reporting the nitty-gritty of the financial details. The company reported Q3 2018 (Q3FY19) numbers yesterday and missed analyst expectations for earnings, revenue, and guidance. Analysts were expecting $3.23 billion in quarterly revenue versus $3.18 actual. GAAP and non-GAAP EPS expectations were $1.71 and $1.93 respectively, versus $1.97 GAAP and $1.84 actual EPS. The consensus for the January quarter was $3.4 billion versus $2.7 billion actual guided by the company. So did NVIDIA drop the ball? Well, let’s find out.

The first thing we need to look at is how the stock performed in a vacuum. NVIDIA reported 1.8% QOQ and 20.6% YOY growth in revenues. The company beat the GAAP EPS consensus by a considerable margin and only fared less than ideally on the non-GAAP EPS. It saw growth in all major segments – except one: gaming. In isolation, the company’s financials weren’t exceptional but they weren’t too bad either.

nvidia-quarterly-trend-q3-2018-1030x534.png


Holistically on the other hand, things start to get dicier. Let’s go over a quick overview of what the bulls and the bears would agree on:

Things the bulls cannot deny:

  • Turing has received a cool reception and the extravagant price tag makes it unpalatable to the pretty much the most of the gaming market.
  • Turing’s new tech isn’t even close to widespread adoption at launch (in other words, it’s unusable).
  • The company’s gaming segment is its lifeblood and a drop in growth there is a serious concern.
  • The cryptocurrency boom (and its positive effects) are over.
  • The stock was trading at a high particularly because of its consistent streak of beating analyst estimates.
  • NVIDIA has a Pascal inventory problem.

Things the bears cannot deny:

  • NVIDIA is in pretty much a monopoly right now and anyone that does not buy Turing will almost certainly go for a Pascal (with the exception of the price point targeted by the RX 580).
  • Turing contains new tech inside of it that has the potential to break the “hardware limits” and give graphics cards an almost magical leap in performance IF NVIDIA manages to implement it successfully (which they haven’t done so – as of yet).
  • NVIDIA has previously priced very expensive dies at a discount just to undercut its competition and part of the reason the company chose to price Turing so high was so they don’t cannibalize their own Pascal inventory. This is one of the reasons why the company beat GAAP EPS but not non-GAAP EPS.
  • Turing is a success in the professional segment, where the Tensor cores can go head to head against offerings from Google for the first time.
So the next question then becomes, who got it wrong? Were the analyst expectations unrealistic or did NVIDIA drop the ball?

In my humble opinion, analysts expectations were widely off center. We published a special pre-earnings piece talking about the reasons that could make NVIDIA screw up this quarter – and we have never done so in the past. Most of the analysts in the know knew that Turing had not received a warm reception and had tapered their expectations accordingly. For most of the analyst estimates it seems like things were still very rosy, Turing was doing great, and the cryptocurrency boom was still raging.

Just like $300 price point for NVIDIA is not justified till it manages to deploy and implement Turing RTX and AI tech, the $165 price point is also not justified in our opinion simply because the company practically owns the market right now. It could go up but at the same time we must express caution. Turing could have the effect of changing mindsets to lengthen the historical upgrade cycle of gamers which would shrink the overall size of the market. This would obviously be to the detriment of NVIDIA. There is no alternative to NVIDIA right now.

It’s certainly feasible that this is a buying opportunity. As we noted previously, the fundamental performance of the company has seen growth. Returns may take a while to materialise however and of course, you can be right all day but if the market thinks you are wrong, get the **** out of the way. This estimate is of course ceteris paribus and assumes financial conditions of the global economy stay the same among other things.

We will also tell you the key signal to watch out for: Look for when the company cuts Turing MSRP across the board. That is when they have either written off Pascal inventory and are ready to milk Turing for all its worth or have actually exhausted it and are about to do the same. This could lead to an improvement in earnings more in line with analyst expectations but we urge readers to keep in mind that Turing is a very large and expensive chip. Cuts in price will obviously eat into margin although there is probably sufficient margin at the moment to afford them and start shifting volume.

TL/DR:
  • NVIDIA overpriced cards with technology touted as the next generation of graphics and that developers have done basically nothing to integrate and performs like utter shithole when turned on
  • Supplies of Pascal cards are starting to dry up
  • Mining boom is over
  • NVIDIA stock is taking a shit
This also just in, water is wet and the sun shines. Not sure why this was a surprise to anyone, literally every analyst saw this coming months ago. Excess Pascal stock from the mining boom, Turing cards overpriced to help alleviate that, and now that Pascal is drying up and people don't want to fork over the cash for next-gen cards with only moderate performance gains and next-gen tech that is barely supported and operates like dogshit, NVIDIA is wringing their hands in worry.

I'd say this is a good chance for AMD to catch up and redeem themselves, but they're too busy re-re-releasing old cards with shitty build quality/design/awful performance and sniffing their own farts to do anything meaningful.
 

MeatRokket08

MORTIS
kiwifarms.net
Joined
Dec 20, 2016
While amd keeps releasing the same card slightly occed over and over.
They had some initial success with the Vega 56 and 64, seeing as how both competed well against the 1070 and 1080, respectively. Unfortunately, lack of proper support and too much time in between updates put a stop to that momentum. It's a shame, because you can see the potential in these cards.

Also, AMD had just released a new card, the RX 590 which competes well against the GTX 1060 (I think it even surpasses the 1060 by a non-negligible margin). I'm hoping this is just a footpath to a more wide released series of cards. The GPU market could use some more competition right now.
 

Pickle Pirate

Pickles are just undead cucumbers.
kiwifarms.net
Joined
Dec 19, 2014
Also, AMD had just released a new card, the RX 590 which competes well against the GTX 1060 (I think it even surpasses the 1060 by a non-negligible margin). I'm hoping this is just a footpath to a more wide released series of cards. The GPU market could use some more competition right now.
The RX 590 is a rehashed 580 with smaller nanometer processing, and the 580 itself was a rehash of the 480. It’s barely neck and neck with the 1060 6gb, and has build quality and power consumption issues (not to mention it’s more expensive than the 1060). It is not a competitive card, it’s a bandaid stop-gap measure until they can produce something better, and they’ve given no indication they even have anything in the works.
 

Tackleberry

VERY cool.
kiwifarms.net
Joined
Dec 5, 2017
Zero surprise. ZERO. NVIDIA's been overpricing their cards for 6 or so years; I used to buy NVIDIA almost exclusively for well over a decade until about 4 years ago. When your mid-range cards are running $400, and AMD cards with equal or superior performance are selling for half the price? How can your stock NOT crash? NVIDIA got cocky and started selling their name rather than a solid product.
 

Nazi vegeta

kiwifarms.net
Joined
Sep 3, 2017
They had some initial success with the Vega 56 and 64, seeing as how both competed well against the 1070 and 1080, respectively. Unfortunately, lack of proper support and too much time in between updates put a stop to that momentum. It's a shame, because you can see the potential in these cards.

Also, AMD had just released a new card, the RX 590 which competes well against the GTX 1060 (I think it even surpasses the 1060 by a non-negligible margin). I'm hoping this is just a footpath to a more wide released series of cards. The GPU market could use some more competition right now.

Thats what I meant with releasing the same card. 590 is a die shrinked and occed 580, which itself was already a re-release of the 480.
Vega has been damaged mostly by its higher price at launch but it is indeed a good card.
 

Rei is shit

kiwifarms.net
Joined
Oct 23, 2017
The raytracing stuff is actually really exciting but clearly not ready for prime time as of yet. Also selling it to gamers when they just look at stuff like screenspace reflections and AO and can't tell much of a difference doesnt really help.
 

Pickle Pirate

Pickles are just undead cucumbers.
kiwifarms.net
Joined
Dec 19, 2014
Real time raytracing is still in its infancy, and if you’re not in an area with reflective surfaces it’s doing nothing for you. It’s a nice idea and once they fix the kinks and come up with hardware that can reliably run it at acceptable frame rates, it’ll be a nice graphics option. But setting it up as the end-all-be-all of your next product line when it’s clearly not ready for prime time is pants on head stupid, especially if you’re charging a premium for what is essentially a beta experience.
 
Last edited:

Smaug's Smokey Hole

Brain open for rental(rent free)
kiwifarms.net
Joined
Sep 7, 2016
Thats what I meant with releasing the same card. 590 is a die shrinked and occed 580, which itself was already a re-release of the 480.
Vega has been damaged mostly by its higher price at launch but it is indeed a good card.

I think Vega suffered from manufacturing constraints relating to HBM, AMD seemed to be aware of this and even before manufacturing of the retail product started they had been buying and stockpiling as many HBM modules they could get their hands on for half a year or so. The cards themselves seemed to sell, they were never in stock anywhere, I just don't think they could get that many out to retailers.

AMD must have done really well overall, Vega and almost all RX cards were sold out even though retailers had raised the prices to insane levels - at least where I live. The pricing on their whole lineup was affected, even at the low end. The RX 550 that competes with the GT 1030 cost the same as a GTX 1050, the RX 560 was priced between the GTX 1050TI and GTX 1060 3GB etc.
Things have calmed down but even then right now I'm looking at a store where the 4gig RX 570 costs as much as the 1060 6GB.


edit: For AMD's future they have the near mythical "Navi" coming up, but there will be a Vega refresh and more Polaris before that happens. I think Intel poached their chief architect for Navi recently, that might not be so good considering some people talked about Navi "being his baby".
 
Last edited: