https://nvidianews.nvidia.com/news/nvidia-updates-financial-…
NVIDIA today updated its financial guidance for the fourth quarter of fiscal year 2019, reflecting weaker than forecasted sales of its Gaming and Datacenter platforms.
This is not a good report.
Andy
ouch, 2.2 billion from 2.7 which was from 3.4 billion. I don’t think I’ve ever seen a company swing and miss by 1.2 billion. The hurt is being applied. My first thought was , “Cyclic companies will cycle”
Turing is a cool architecture with cool ideas, unfortunately it is just too expensive. As someone who buys these things I’m glad the gaming market told nVidia to F’ off on the price raises. The Nvidia 2060 was particularly galling as the x060 products have historically been nvidia’s mainstream model. In 4 short years their mainstream model has increased from 249 dollars to 349 dollars. THe worst part is the 2060 doesn’t even have enough power to run any of the raytracing stuff at acceptable speeds. On top of that it is severely hamstrung by its lack of ram.
I look forward to the data center updates, Glad I’m out of all data center hardware stuff. Winter has come!
-e
It’s interesting to me that this happened. People on the board that I really respected believed greatly in Nvidia, and were sure it would conquer the world, and I was convinced three times to take try-out positions in it, but I just couldn’t hold it and sold out each time fairly quickly. They were rolling along at the time and I didn’t have a good reason, except huge market cap compared to our other companies, it made hardware instead of software, its revenue wasn’t recurring, my bad experience in the past with other chip manufacturers, etc. But I guess the overwhelming factor for me was that I was simply not comfortable holding it in spite of the excellent numbers it was putting up. What can I say?
Saul
They were rolling along at the time and I didn’t have a good reason, except huge market cap compared to our other companies, it made hardware instead of software, its revenue wasn’t recurring, my bad experience in the past with other chip manufacturers, etc.
I don’t think we can stress the software model enough. The simple reason? Gross Margins. Software GM’s are typically higher, but just as importantly, they’re also MUCH less prone to fluctuate wildly, because the product mix doesn’t really change. There will be new versions, development, etc, but the gross margin will remain about the same.
Let’s look at where things can go wrong with companies that sell physical products:
-market for a product reaches saturation (subscription solves this)
-new product doesn’t sell like the old one did (switching costs largely solve this)
-new product components are more expensive and this cost can’t be passed on to customers (not typically an issue with software)
-customers don’t like the new product (switching costs certainly help here)
-competitor innovates a better product (switching costs certainly help here)
-etc
With software, especially subscription software, things are so much more stable. Software companies can of course get out-innovated, but they have to really drop the ball – after all, they got their customers for a reason. As long as they keep up with the latest industry trends and keep customers happy, almost always they do just fine. And as we’ve seen, companies that are really changing the playing field can grow massively quarter after quarter.
Bear
Well I was building a position in NVda thinking that all they had to do was move past the inventory glut from crypto and it would all be smooth sailing and gravy.
Saul reminded me this morning that I also have never been very good at timing semiconductor companies.
Today’s miss seems much more about overall slowing in China and Taiwan.
So I’m unwinding my position, already sold 55% of my shares at 138.80 for a 9% loss.
I think I’ll stick with the reacurring revenue model from now on. 
Chris
I did get a capital loss off Nvidia late last year. That sucked but was worth real dollars. It goes to show that semiconductors, any semiconductor, is cyclical and thus do not deserve a higher multiple although they can go for years as if they rule the world.
I look at the AI companies looking to create AI chips on the edge. I look at Pure trying to get its product as part of the edge. But you know what, Nutanix or VMWare will be part of almost everyone’s edge no matter what chip or server/storage device they use on the edge as the edge will become just another part of the data center (and please, we have had that argument and I was correct, data centers on the edge using cell phone towers to locate).
I look at Nvidia these days because it is such an interesting company. As an investment (although at some point things will return as autonomous vehicles and machines start to roll out - may take longer than we think - but that will be big) Nvidia missed by $700 million in a quarter on their core product without even giving a warning. Talk about no insight to their own sales channel. They now are warning (my glance may be wrong) but it seemed like close to $1 billion less, maybe $500 million less, a huge number less than even their decreased guidance at last earnings call.
China is a big culprit here. The software and SaaS companies we follow have little business in China. It should not impact them at all. But it is not possible to invest in a company whose executives appear to have such little insight into their end markets.
Make no mistake about it, at some point even SaaS companies will slow, but that is not today or tomorrow. And it will not be because of China.
So continue to follow Nvidia I shall as they are an interesting and important company dictating the future of AI that is so important. But as an investment cheap keeps getting cheaper.
Tinker
I was one of the founders of a fabless semiconductor company years ago, and in addition to the problems Bear mentioned above, for all but the biggest fabless chip companies the supply chain also can be a big problem – as it was on occasion for us. Unless you are a top tier customer at TSM etc., you could have trouble reacting to rapid demand fluctuations (in either direction) and also face burdensome buffer inventory requirements from OEMs that are incorporating your products in their systems.
Clearly this is not a problem for SW companies.
Rich
I don’t think we can stress the software model enough. The simple reason? Gross Margins. Software GM’s are typically higher, but just as importantly, they’re also MUCH less prone to fluctuate wildly, because the product mix doesn’t really change. There will be new versions, development, etc, but the gross margin will remain about the same.
NVDA certainly has and will continue to have some disadvantages when compared to SaaS companies that provide cloud delivered SW on a recurring basis. However, NVDA has maintained GMs of around 65%. Well, the latest guidance for Q4 will drop the GMs quite a bit. But my point is that GMs are not the issue with NVDA. NVDA’s GMs are (were) higher than TWLO’s GMs (TWLO’s are in the mid-50s).
Chris
I wonder if some of NVDA’s market share stolen by XLNX.
XLNX revenue was up 30%+ y/y and stock reflected that by huge jump last week.
in addition to datacenter, XLNX also seeing growth partly because of early build out of 5G. Its chips are in every radio (cell tower) thats being put up for 5G around the world.
I dont hold either of these… thanks to Saul and this board, I am happy with cloud / saas stocks!!
XLNX Data Center and TME segment was up 14% and within that advanced systems was up 66%. They are gaining cloud wins for inference.
But the numbers are small. The entirety of the Data Center and TME segment was $168M. So we’re talking about advanced systems sales that have to be at least as low as mid double digits millions to make that math work out. Nvidia nearly had $800M in data center sales.
Xilinx is not winning the training share that Nvidia has. The numbers they are putting up in AI wouldn’t lead to an additional $.5B miss.
I think it’s mainly macro slow down on HPC spend. Over last two years the world has built out an amazing exponential increase in computing power(FLOPs on line). That has benefited Nvidia. I think we are at one of those rare times when the processing power has exceded the consumption. The world has to figure out how to consume all that processing power. Then they’ll start to crave more again. Not good for Nvidia. Worse for all those Nvidia “killing” startups.
After Nvidia’s Q3 it was evident that something of biblical proportions was taking shape. This put cracks in or destroyed some of the core thesis’s with the company. This just adds to the mayhem. The thesis for Nvidia is inarguably damaged and will remain that way for quite some time I think.
Darth
The article below argues that the cloud boom is dying and it will have a slowing effect on the tech industry in general.
Plus many of our tech stocks have been a safe haven in the ongoing trade war. In the event of a trade deal money will start flowing into some of depressed industrial sector companies like Caterpillar.
Could a cloud burst and a trade deal deliver a 1-2 punch to the growth rates and multiples of our software stocks?
https://www.marketwatch.com/story/nvidia-is-latest-sign-that…
Nvidia Quarterly Revenue Trend Q3 FY19 (in millions USD)
$1764 - Gaming
$305 - Professional
$792 - Datacenter
$172 - Auto
$148 - OEM
$3181 - Total
Nvidia is now forecasting $2200, basically a billion less Q over Q, concentrated mostly in Datacenter and Gaming.
Just a wild guess (Q over Q):
$1100 - Gaming -38%
$250 - Professional -18%
$550 - Datacenter -31%
$150 - Auto -13%
$150 - OEM -flat
$2200- Total -31%
IMO, Nvidia is back down to pre-crypto early-2017 levels of revenue. I’d also put early 2017 price on it, $110-$150 range for the foreseeable future.
As a side note, my company is seeing a revenue decrease. We have cut back on several hardware purchases. Yet we are still paying on our subscriptions every month.
Yeah, I’m one of those people . . . They can’t all be winners I guess. Longer term, I expect Nvidia to return to exhilarating growth, but at present and the next several quarters they’ve got to pull back on the reins.
It’s OK to be wrong, it’s not OK to stay wrong . . . And old habits die hard (that hopeful notion that the recovery will make up for the lost capital and lost opportunity).
It’s time to bail and put the money to use elsewhere.
I’m tempted to buy more IIPR. The pot business is in the early innings and no one knows who the winners are going to be, but IIPR doesn’t much care, they are selling (leasing, actually) infrastructure that supports the industry, even if one of their clients goes out of business, they will have little trouble leasing the facility to someone else. And with the inevitable decriminalization of pot, I imagine IIPR will explode upward.
Yeah, I know, another physical product with all its inherent problems, but this is the front of a huge wave that will be cresting for years.
Or maybe I’m just an old stoner . . .
Nvidia Quarterly Revenue Trend (in millions USD)
Q3FY19 Q4FY19
Gaming 1,764 954 -46%
Professional 305 293 -4%
Datacenter 792 679 -14%
Auto 172 163 -5%
OEM 148 116 -22%
Total 3,181 2,205 -31%
Nvidia meets reduced revenue target. Gaming was worse than expected, Datacenter not as bad as expected. Guidance revenue is flat y/y. A/H price is up 5% to $162, already out of my price bracket/