NVDA Q319

Hi Greg,

Obviously still up a lot YOY, but the recent quarterly trend is pretty precipitous. Personally, my thesis for NVDA revolves around that datacenter growth, and the drop-off is enough to put me on the sidelines.

I agree with you. The Datacenter growth is what had me the most excited also. When I first looked at NVDA it was growing really strong. But I do not think that thesis is broke, I just think it took a breather because of the new Turing chips coming in. We should know by next quarter whether that is true or not.

Now I am not really that excited about the automotive but it could lead to some great gains in the future but the new Rapids launch for Data analytics and machine learning really has me excited.

I can understand people dropping this and deciding to get out but I see to many good things coming so I will just hold on. Next quarter isn’t going to be good but then it should start picking up we will see.

Thanks for the response Greg and your views.

Andy

3 Likes

Hi PB,

My biggest concern going forward (I’ve never owned the stock but been tempted by it) would be that last quarter on their CC they said that inventory (Finished goods and channel) wasn’t a concern, but it turned out to have a massive negative impact. This whole collapse in the crypto business and the ramifications thereof seems to have taken them totally by surprise which is NOT a sign of wise fiscal and operational management.

I agree with you but I am not sure how much blame Huang should get for this. He obviously didn’t think that AMD’s inventory would hurt him and He didn’t think the Crypto business would hurt his gaming business. He was wrong on both accounts. But I remember every time, as a gamer, I took advantage of discounts like this on computer parts, I ended up being unhappy with them and out looking for new parts within 6 months. So all this inventory should be blown out in 6 months at the most. The new Turing card might make gamers want to change faster when they finally see the ray tracing. Thanks for your thoughts.

Andy

2 Likes

Of course, the problem with dropping and expecting to come back is that one has to develop the conviction to come back before the stock has already gained substantially or one is giving up those gains. To drop because one has lost confidence in the company overall is one thing, but dropping with the intention of coming back gets one into timing.

2 Likes

Great, open, honest in depth analysis.
What leads you to the prediction of a double in 3 years, though?

1 Like

Also, NVDA came up with a GPU acceleration platform for data science and machine learning called RAPIDS, which lets companies analyze massive amounts of data and make accurate business predictions at unprecedented speed. Up to now this was all done by CPU’s and this system is 50 times faster. This is for the OEM&IP market. This is what Colette had to say.

The RAPIDS launch opens up a $20 billion server market used for data analytics and machine learning workloads to GPUs, and it’s received broad industry support, including from Oracle, IBM, SAP, Dell EMC, Hewlett Packard Enterprise, Microsoft Azure machine learning, Google, Q-Flow as well as the open source community.

Great point. Just reading over an article I got from ARK investments, they had this to say:

Nvidia’s RAPIDS framework brings the power of GPU acceleration to common machine learning and data analytics frameworks today. Unlike its foray into deep learning, where it created the market from scratch, Nvidia launched RAPIDS which taps into the large and growing market for analytics. With Moore’s Law grinding to a halt, we believe Nvidia’s move could not be more timely.

Isn’t this in our favorite Alteryx’s wheelhouse?

Dominic
Long AYX and NVDA

2 Likes

Repost from NVDA board:

Thanks. That was a great post. You may be right. This could be a one quarter problem, and NVDA will probably get back on track. I think the $24,000 question is “Back to what?” No one outside and possibly inside NVDA seems to be able to accurately forecast NVDA’s chip business. Certainly, NVDA will be no worse than medoicre, but more likely a good to fabulous company in the future. The problem is there is no way to model anything.

Although I believe in holding companies as long as possible, I don’t believe in holding companies that have a high probability of short term pain and an unclear future. I realize that many Fools believe in holding through anything, which will probably work in this case. I will be hopping on the sidelines and watching. I might re-enter after a quarter or two. Although market timing does not work, avoiding a period of pain for an individual stock often does. I share everyone’s pain. Best of luck to longs.

bulwnkl

6 Likes

What leads you to the prediction of a double in 3 years, though?

Hi FlyingCircus,
That is a great question and one that everyone needs to think through. I always take predictions with a grain of salt because nobody is really good predicting the future out for any amount of time but here are my thoughts and in 3 years we will see if I am right. I will keep this post on my NVDA writeups to measure NVDA against.

NVDA is sitting at $164.43 and it’s marketcap is $99.973B. It’s P/S is 8.42 and it’s P/E is 23.97.

Right now they are growing at 21% and this was a terrible quarter, In order for them to double in 3 years they have to have a Compound Annual Growth Rate (CAGR) of 25.9%. So for my prediction they have to grow 4.9% faster than last quarter. They have been growing the revenue reliably since the Q217 quarter at 24% and then accelerating. Let me show you the Revenue Growth numbers.


Q117    Q217    Q317    Q417    Q118    Q218    Q318    Q418    Q119    Q219    Q319
13.04%  24.35%  53.85%  55%     49.23%  55.94%  32%     34.10%  65.46%  39.91%  20.45%

Now this is backward looking but in 2017 they had an average cagr of 36.5%. In 2018 they had an average Cagr of 42.82% and in 2019 I am going to give Q419 a growth rate of 20% so the average growth rate for 2019 will be 36.75%. After the Q419 quarter I expect them to start accelerating because the gaming inventory glut will have been worked through. The new Turing card with ray tracing will be the gaming card everyone will want, because more games will support it.

They have the new Turing card for the Data center’s which should start growing because in their conference call they claimed the Turing T4 cloud GPU has received the fastest GPU adoption of any server GPU. They also announced an upgrade on their TensorRT stack which will allow multiple models and frameworks to be run.

They launched a new NVDA RTX server for the Professional Visual team which will take market share from the dominating CPU server business. This will be a small increase that shouldn’t be looked on as a huge business.

They launched Rapids, which is for data science and machine learning. I think this will be a bigger (trying not to be to excited) revenue increase and is a new business they are going into.

Finally, automotive is starting to ramp up with their new level 2 system which over the next 3 years should be growing to probably a level 3 system and maybe a level 4 system.

So flying circus that is why I think they can double in 3 years, and also I can’t even imagine what other business’s they will find for the use of their chips. I think everyone has to remember one thing. In the Gaming business it was the middle of the road Graphic chip that is having problems due to over supply and cost conscious gamers. The Upper high-end chips did fine. Also, with the Data Centers, and why they were light. They had a new chip coming out and I think it just took a breather. If anyone thinks AMD is taking market away from NVDA on the High end market I would like to hear their thoughts. But with the new Turing card, and NVDA providing that card to the middle market probably by the end of this quarter, I think AMD will be hurting again in the gaming market.

Andy
If anyone has problems with my thoughts or with any of my math please speak up. Also I am not trying to convince anyone of staying or getting into this stocks, just enjoy the discussion of companies and stocks.

11 Likes

Hi Bulwnkl,

Although I believe in holding companies as long as possible, I don’t believe in holding companies that have a high probability of short term pain and an unclear future. I realize that many Fools believe in holding through anything, which will probably work in this case. I will be hopping on the sidelines and watching. I might re-enter after a quarter or two. Although market timing does not work, avoiding a period of pain for an individual stock often does. I share everyone’s pain. Best of luck to longs.

I can’t fault you for doing this but I just see this company as a great company. The loss has already been taken and although next quarter will probably be flat to slightly up, I am not sure what the market will think about that but it could go down further. Normally I would probably move on also but I think this company will double in 3 years so I want to hold this company. Any company I can get a Cagr of 25% on YoY is something I just hold and tuck away. But don’t worry I will watch it for you and let you know when might be a great time to get back in :).

Andy
That is what these boards are for.

8 Likes

Andy,

If anyone has problems with my thoughts or with any of my math please speak up.

My only problem is that you are using about 26% CAGR for the next three years (the necessary number for double in 3 years). They have been growing at 33% to 35%. If you visualize that 2018 just sucked revenue out of 2019 the result is 33 to 35. Use 33% CAGR and your double could come a quarter sooner.

KC

1 Like

Hi KC,

My only problem is that you are using about 26% CAGR for the next three years (the necessary number for double in 3 years). They have been growing at 33% to 35%. If you visualize that 2018 just sucked revenue out of 2019 the result is 33 to 35. Use 33% CAGR and your double could come a quarter sooner.

After listening to all these conference calls over the years I thought the game was to predict low and then when you beat the prediction you get a big up side in the price of the stock :slight_smile:

But in reality the future is still unknown and anything could happen to ruin my prediction. If we go into a recession I think I will be very happy to double in 3 years.

Andy

Andy,
Like you I continue to hold Nvidia. In May after the Q management told us that crypto was just a 10% business. Yet they had very high inventory (>>10%) at the end of each Q. Could crypto have been >10% of their business? This makes me wonder if gaming is really a $1.7B/Q business or something smaller say 1.4B/Q. One analyst asked that question, i.e. next Q your gaming will be about $1.1-1.2B, so will you start to grow from that base or not but management deflected. Other than that I believe DC, auto, visual will continue to grow and this company should have a double digit CAGR over the next 10y.

2 Likes

A short article specified that CEO Huang had recently sold $40 million worth of shares. Perhaps he did. However, on the insider roster on Yahoo, his last sale was at $166 for around $3 million. He has automatically been selling, periodically, all the way up, with “small” amounts such as that.

I did not see this great insider selling by Mr. Huang that a short article focused on as an aspect to aggravate the recently issues. Perhaps someone else has found it. It would need to be a large insider sale within that last 3 months, subsequent to the previous earnings call just prior to the one a few days ago.

I still think it seems difficult to believe (albeit, not impossible) that a $700 million miss ($29 million - relatively small, and happens from time to time) was in the data center, but the rest (and actually more as there was upside to other areas, was all in gaming.

True to form Nvidia did very well in the high end again with the new Turing chips, and this should only improve as more games come out, and in particular in 2020 when the next upgrade comes out to fully cement the new architecture.

I do not necessarily think that AMD was taking more market share and commoditizing the market. More likely that NVIDIA GPUs (I have the 1070, which is the minimum necessary to operate my VR system, and it was something like $500 less than moving up to the 1080 in regard to a Dell gaming desktop). That was not an insignificant difference, and for me the 1070 was find. Crypto miners were probably selling a lot of Nvidia chips as well.

The issue with Turing, if I understand it correctly, is the chip is clearly more powerful, but if one incorporates ray tracing in the software the FPS does not increase all that much. Which of course makes sense as ray tracing takes an incredible amount of computation, and has never before been enabled on a gaming desktop before because it is so computationally difficult to do. Nvidia has done it with a combination of hardware and software, including AI to literally think ahead to take some of the computational requirements off the chip.

Thus, from a consumer perspective, you need to start seeing the new games to show the clear advantages of using ray tracing in order to pay the premium. And to date the new high end Turing chips have quickly sold out. But apparently not at great enough volume to offset the mess that otherwise exists and it is remarkable that Nvidia had so little insight in regard…practically malpractice.

But I will get off that issue. Just wanted to get into this insider selling thing as it besmirches Mr. Huang if he suddenly made a much larger insider sale in the $40 million plus range at just the time the business was, temporarily at least, falling off a cliff in gaming chips. I did not see any such sale in Yahoo! Does not mean it does not exist, but it is not there anyways.

Tinker

2 Likes

A short article specified that CEO Huang had recently sold $40 million worth of shares. Perhaps he did. However, on the insider roster on Yahoo, his last sale was at $166 for around $3 million. He has automatically been selling, periodically, all the way up, with “small” amounts such as that.

Looks like about $11.2M sold between Sept. 19th and 21st (41,253 shares at an average price of $271.98):
http://d18rn0p25nwr6d.cloudfront.net/CIK-0001045810/f606569d…

https://investor.nvidia.com/financial-info/sec-filings/defau…

1 Like

There goes any trust that Huang did not at least have a good idea of what was happening in the channel. Prior to this he had not regularly made large insider sales. Instead he made a few million here, a few million there, etc., in a systematic regular selling process.

Tinker

Mr Ferragu from New Street Research came on CNBC this morning and stated that Crypto gave NVDA an additional $9 million dollars of additional revenue in the last 4 quarters but that $450 million dollars of inventory was stuck in the distribution channel for Crypto and that is what NVDA is trying to expel from the channel.

Andy

2 Likes

I think its hard to predict what growth rate NVDA can do, because we don’t know what the gaming growth rate was without Crypto, so we can’t use the past growth rate to predict. Management doesn’t know either.

At the 2nd Quarter conference call they said inventory was fine, they were not going to plan for the $100M of crypto sales going forward. Turns out the crypto sales were much higher than that and part of the gaming sales where going to crypto and they lower their forecast next Q by $700 M.

So now we have the growth rate going from 40%—>21%----> to -7% next quarter.

Here are the questions I’m asking myself on NVDA.

If management just missed their gaming center demand by $700 M last quarter, how confident can I be that they have the inventory problem solved at the end of this quarter and it won’t carry forward?

Data center demand missed forecasts. Is that caused by crypto also?

What is the real growth rate for gaming, and the company, without crypto?

When is the auto business going to kick in and become another growth driver?

I can’t answer any of the questions above, so I can’t determine how likely it is that they grow 25%+ going forward.

NVDA is clearly a great company, with a management team that has been great in the past but just stubbed their toe pretty bad.

Jim

2 Likes

Andy,

but that $450 million dollars of inventory was stuck in the distribution channel for Crypto

I think that this is what I essentially said in my post on NPI. The real growth was 33% or $9.290B versus $9.714B reported. Difference is $424 million. So if $9 million real crypto sales, then $415 million stuck there. So if gaming was really only $5.19B and there is $0.415 to $0.45B stuck there, how long to clear it. These were for desktop gaming. Don’t know what % of gaming that is. And again, the unknown AMD inventory.

KC

1 Like

A short article specified that CEO Huang had recently sold $40 million worth of shares. Perhaps he did. However, on the insider roster on Yahoo, his last sale was at $166 for around $3 million. He has automatically been selling, periodically, all the way up, with “small” amounts such as that.
--------------------------------------------------------------------------------------------------------------------

Looks like about $11.2M sold between Sept. 19th and 21st (41,253 shares at an average price of $271.98):

That was a very useful link you included, linking to the SEC form 4. Based on that form he has over 20 million shares in total in various accounts, and he sold all of 41 thousand, which comes to about 1/500th of his total, or a massive 00.2% of his position.

Now, I’m not invested in Nvidia at present, but if you are a stockholder, I think getting all worked up about that insider trade is silly on the face of it.

Saul

22 Likes

That was a very useful link you included, linking to the SEC form 4. Based on that form he has over 20 million shares in total in various accounts, and he sold all of 41 thousand, which comes to about 1/500th of his total, or a massive 00.2% of his position.

Now, I’m not invested in Nvidia at present, but if you are a stockholder, I think getting all worked up about that insider trade is silly on the face of it.

Saul

Agreed.

Huang’s paper net worth (only including his NVDA share holdings) has dropped from a peak of ~$5.8 billion in August down to about $3 billion at present, so his interests are definitely aligned with a high NVidia share price.

-volfan84
still long NVDA ($17.52/share cost basis)

1 Like

Now, I’m not invested in Nvidia at present, but if you are a stockholder, I think getting all worked up about that insider trade is silly on the face of it.

Saul

Insiders sell shares for a variety of reasons. IMO it is usually of no consequence. It is when insiders buy shares that my interest is piqued.