We called out in a report Nov. 12th , ahead of Nvidia’s (NASDAQ: NVDA ) earnings, that there was risk. But after seeing the numbers I think there’s some bigger issues than just crypto channel inventories. I think their mid-priced RTX is going to have issues selling through. And Datacenter seems to have some issues.
I think it’s very important to understand a company’s mix of business. As hyped-up people get about Datacenter and Auto, if Gaming isn’t going to do it, you have a dead stock for a while.
Gaming is almost 60% of the company. Auto is tiny and hasn’t been moving much. And Datacenter has been slowing, maybe for some fundamental reasons.
Let’s review.
Here’s What We Said Ahead Of Earnings Here’s what we said in a report last Monday ,
“But there are some near-term headwinds.
Channel inventories and a general video game slowdown in the US and China make this quarter and the guide a little tougher. Who doesn’t want to own Nvidia? But this does not seem like a great time for an entry.”
Here’s what we told subscribers in September when we downgraded from Buy at 275 ( pay wall ),
“I think the right October gaming growth can be low singles (percent sequentially) telling you that 10%-40% of sales growth are cut off by dropping cypto demand and too much inventory in the channel. I think Q4 is at risk to this too.
The actual gaming growth sequentially for the quarter just reported was -2%, a big miss, even for us.
It’s Not Just Crypto Our note last Monday pointed out that reviews for the new RTX either complained that the high-end product is just way too expensive or the 2070 mid-priced product isn’t worth the upgrade.
But here’s what CEO Jen-Hsun Huang said on Thursday’s earnings call:
“It’s all about crypto hangover.”
He was saying that the only reason for the miss is crypto inventory.
I’m not so sure that’s the only problem.
We have no problem that the 2080 series is selling great, but that’s the high end of the market. As we said Monday the high end is not big enough to drive the total pie for Nvidia.
Over the last couple of years Gaming made up 55%-58% of overall business and 60% of the back half.
They need to nail Gaming.
You can read the negative industry reviews in our last report last week but I’d like to focus today on the ray-tracing reviews.
Let’s take a quick step back.
Some of the slowdown of demand may have been gamers waiting for the RTX so that slowed channel-sell-throughs.
Some of the slowdown though may be that there’s a limited amount of ray tracing games available on the market. The thinking would be that when you have more choice of ray tracing games people will buy the GPUs.
But there’s also a problem that even when we have the ray tracing games, it still may not be worth it to upgrade.
If that’s the case, this problem was not just crypto.
Let’s see what Huang said on this:
“And today with Battlefield V, people are enjoying real-time ray-tracing for the very, very first time.”
This is a must, right? If people are not “enjoying real-time ray-tracing” then Nvidia has a mass market problem. The mass market, which is most of their business, has to love that product.
That’s as opposed to high-end gamers who will probably spend anything on a GPU to compete.
Let’s see if they will “enjoy real-time ray-tracing” and let’s see the reviews of Huang’s call-out Battlefield V with ray-tracing enabled.
Here’s what Tom’s Hardware had to say specifically about ray tracing for Battlefield V.
“There’s also the question of if the game will look good enough to justify lower frame rates and a more expensive graphics card. But the game is available now, and the effects are mixed, at least to our eyes.
…but when playing the game’s campaign or multiplayer while trying not to get shot, higher frame rates and resolution offer more benefit to the game than seeing light from an explosion bounce off the side of a tank.”
What Tom’s Hardware was saying is that if you turn on ray tracing your game play gets hurt. You trade ray tracing for framerates. Do gamers want to lose their competitive edge in favor of seeing special shadows and reflections? Probably not.
This may pose a problem to RTX’s initial mass-market appeal.
Here’s what Extremetech said specifically about Huang’s call-out, ray tracing for Battlefield V.
“.
.but gamers are essentially being asked to fork over even more money to play games at lower performance levels.
…But regardless of how you slice it, gaming in RTX imposes a substantial penalty, and only the most well-heeled gamers are going to be playing in resolutions above 1080p. The performance impact of RTX looks to be heavy indeed.
”
So, what’s your take?
Mass market is what is going to drive Nvidia’s Gaming line-item which is most important to Nvidia’s earnings and stock price. That one line-item is the most important to the company and its investors. That one line-item is solely dependent right now on RTX doing well in the mass market.
We know the high end will do well, it’s already doing well.
Nvidia needs that mass-market mid-priced product to win, but gamers are asked to give up performance to turn on RTX and have more impressive shadows and reflections. It doesn’t sound like it will be worth it to upgrade.
So even when we have enough ray tracing enabled games it may not be a Pascal-type boost for Nvidia.
I think there’s no doubt ray tracing is a leap for gaming.
But if it comes at the expense of game play or framerates , as powerful the GPU is, gamers may take a pass.
So when we clear this Pascal channel inventory you then need the mass consumer to want RTX, at a higher price. Will they? I think that’s very much in question.
So far the one main ray tracing game is not worth it to turn on RTX.
Here’s what Huang said about the performance of RTX,
“RTX is higher performing at the same price point than any graphics card on the planet. And so at every single graphics – every single price point, it’s the highest performing graphics card.
So it’s unambiguously the highest performance GPU in the world and then – and of course, all of these great new features will be coming.”
But what I’m surprised that no analyst asked on the earnings call is that maybe so much power is being used up on the graphics element that steals from the game play. So, yes, you have a higher performing card, but for something other than enhancing your gameplay.
How Much Inventory Is In The Channel Nvidia said on Thursday’s call:,
“I think the channel has more than 12 weeks of inventory between us and the other brand. One of the things that’s hard to estimate is how much inventory the other brands have.”
Twelve weeks is a full quarter of inventory.
Nvidia needs to stop shipping for a quarter to get back to normal.
And how much does the “other brand have?” AMD (NASDAQ: AMD ) told us on their recent earnings call,
“And so, I would view that – we need to work off some of this channel inventory that’s in place, and then go back to sort of a more typical seasonality, which would see the second half stronger than the first.”
AMD implies they don’t expect inventory to clear until the second half of next year. That means they will not have worked down the excess channel inventory until Q3 2019.
AMD is telling you there’s a ton of inventory in the channel.
Here’s what Nvidia said on that.
“And we really didn’t see that until towards the end of the quarter.”
Nvidia was blindsided about this inventory problem.
Thursday’s negative earnings call was something Nvidia only realized “towards the end of the quarter.”
But Nvidia said Thursday: “It will take one or two quarters to work through” that channel inventory. But AMD is saying it will take three quarters to work through it.
Who’s right?
Personally it looks like AMD has this more under control.
Look at AMD’s inventory growth.
2017 2017 2017 2017 2018 2018 2018 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Inventories 839 833 794 694 715 750 738 YOY 24.3% 12.1% 2.8% -7.6% -14.8% -10.
0% -7.1% Source: Elazar Models, data pulled from earnings reports
AMD’s inventories are fairly clean down 7% year-over-year.
Let’s compare that to Nvidia’s inventories.
2018 2018 2018 2019 2019 2019 Q2 Q3 Q4 Q1 Q2 A Q3 E Inventories 855.
00 857 796 797 1090 1417 YOY Growth 64.11% 26.22% 0.25% -2.92% 27.
49% 65.34% Source: Elazar Models, data pulled from earnings reports
You can see Nvidia’s inventories jumped 65% year-over-year, meaning they were blindsided by this channel issue.
Their guide is for revenues to be down 15% next quarter but their inventories are up 65%. Nvidia has too much in-house inventory as well. That hurts margins when they have to slow production because they have too much inventory.
And while a company’s own inventories are not channel inventories, they can be reflective if there’s a build up in the channel and to what degree.
It looks like Nvidia’s inventory situation is much worse than AMD’s and AMD is saying it’s going to take three quarters to clear. I’d prefer to follow that line of thinking.
So if Nvidia is wrong about it only taking 1-2 quarter to clear you have further gaming risk in Q4 and Q1.
And that’s just talking about channel inventories. If their mass-market product does not do well you add another problem on top of that.
There’s further risk in 60% of Nvidia’s business and it’s not just due to crypto.
What’s Up With Datacenter Datacenter was also disappointing.
Here are the numbers.
Calendar 2015 2015 2015 2016 2016 2016 2017 2017 2017 2017 2017 2018 2018 2018 Fiscal 2016 2016 2016 2017 2017 2017 2017 2018 2018 2018 2018 2019 2019 2019 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 Q4 Q1 Q2 Q3 A Jul Oct Jan Apr Jul Oct Jan Apr Jul Oct Jan Apr Jul Oct Datacenter 73 82 97 143 151 240 296 409 416 501 606 701 760 792 Datacenter QTQ Growth -17.0% 12.3% 18.3% 47.4% 5.6% 58.
9% 23.3% 38.2% 1.7% 20.4% 21.0% 15.
7% 8.4% 4.2% Source: Elazar Models data pulled from earnings reports
The problem here is that (you can look back) the October quarter always accelerated from the July quarter, seasonally.
That was not the case in this just reported October quarter.
An analyst asked on the call if they can do 20% in the fourth quarter. I think they just showed you they are underperforming normal seasonality and we should not expect much here either.
When asked about a key driver “inferencing” here’s what Huang answered:
“The ramp of T4 is completely related to customers porting their model on top of our platform. And the inference model is really complicated. This is one of the things that I’ve talked about in the past that on the one hand, people think that inference appears to be simple because there are so many ASICs built being talked about.
And when that happens, it comes down to their decision of how many they would like to buy, and that tells us about our adoption rate.
…And I look forward to coming back and telling you guys about success.
“Complicated” and “when it happens” are not something you want to hear about a main driver for datacenter. “When” is something that’s nothing happening and is not a confident term.
Machine learning had been their main driver ahead of 2018. Now it’s supposed to be inferencing. Something’s up.
Here’s what Intel (NASDAQ: INTC ) said on their datacenter business:
“.
..and our Data Center business is growing at more than twice the rate we expected in January.
…but outstanding year this year for Data Center, both top line and bottom line.”
Xilinx (XNLX) also had bullish comments on datacenter :
“Establishing ourselves in the data center and accelerating growth in our core markets resulted in a new quarterly revenue record of 746 million, which is up 19% year-over-year and drove the non-GAAP EPS, up 30% year-over-year.
Here’s what AMD said on datacenter :
“Customer interest in the product is strong based on its performance and differentiated feature set, and we have already secured multiple datacenter wins with shipments expected to begin in the fourth quarter.”
Overall it sounds like there’s more competition coming in datacenter which is showing up in Nvidia’s slowing growth rates along with their between the lines commentary that “inference” is “complicated.
”
Intel has owned inferencing. Besides the added competition in the space, penetrating Intel’s bread and butter may be a little tougher than Nvidia’s prior experience.
Datacenter also has nothing to do with retail channel crypto inventories.
Conclusion Even though the company said the problems were just crypto, I’m not so sure. You have datacenter slowing sequentially now for three straight quarters vs. competitors doing great and launching new products.
But the real key for Nvidia shareholders is how well RTX will do at the mass market. It sounds like things have stalled there because RTX may not be worth the upgrade yet even when ray tracing games hit the market.
We published the above to give a taste of what Nail Tech Earnings members receive daily on multiple companies. We’re speaking to a ton of companies and modeling to see who can blow out earnings this year.
We’ve had nicely positive returns in this market finding great tech stock buys and finding the right time to get big.
Dip your toe in the water with us with a free trial .
Wishing you a ton of success!
Disclaimer:
All investments have many risks and can lose principal in the short and long term. The information provided is for information purposes only and can be wrong. By reading this you agree, understand and accept that you take upon yourself all responsibility for all of your investment decisions and to do your own work and hold Elazar Advisors, LLC, and their related parties harmless. All model portfolio trades are hypothetical to show direction, conviction and timing. Performance excludes all relevant transaction costs. Elazar and its employees do not take individual stock positions to avoid front running and other potential customer related issues.
Disclosure: I/we have no positions in any stocks mentioned, and no plans to initiate any positions within the next 72 hours.
I wrote this article myself, and it expresses my own opinions. I am not receiving compensation for it (other than from Seeking Alpha). I have no business relationship with any company whose stock is mentioned in this article..
