NVDA 2nd quarter

I do not find the price unacceptable at all after reading the CC and in terms of my own ratios which include fundamental qualities of the company, even if undeniable risk does lie in the assumption, possibly gratuitous, that Nvidia is the clear leader in the field and will remain so for the foreseeable future (I was never interested in the crypto red herring.)

Despite the risk of challenge, I will definitely increase my holding, currently quite small, if there is a really good old-fashioned fall in the price. Few enough quality companies come through my screens and even a chip stock might be worth a serious flutter. But a flutter it is; the old ambush and obsolescence story.

4 Likes

Another great quarter, with beats on earnings and revenue. Guidance was for another record revenue quarter, but short of estimates by $90M. Stock is down 4%.

Earnings $1.76/share

  • beat estimates of $1.66
  • up y/y from $0.92
  • gross margins 63.3%, up from 58.4% y/y

Revenue $3.12B

  • beat estimates of $3.10B
  • up 40% y/y, from $2.23B
  • down 3% q/q - seasonal

Guidance $3.25B

  • lower than estimates of $3.34B
  • due to low crypto demand

Segments

  • $1.805B gaming up 52% y/y
  • $281M professional up 20% y/y
  • $760M datacenter up 83% y/y
  • $161M auto up 13% y/y
  • $116 OEM & IP down 54% y/y
4 Likes

For Q3, what was ~$300 million in crypto currency is now assumed to be zero, and with the expectations of Turin gaming cards to come out in time for Xmas, GPU gaming card sales will be artificially low for Q3. This is because when a new architecture is announced, the old will slide down the ranking tables and their prices will be discounted to sell in the mid-level, while the new architecture will be selling in the high level.

Would anyone knowing this not wait for the new?

That is hwy Q3 guidance is as it is. And yet still, growth, despite the largest segment of NVDA business taking a quarter off largely in anticipation of the new chip set, a chip set that will largely have a world wide monopoly.

What NVDA also announced is that the new chip set will open up an entirely untapped market in computer graphics that is enormous. It is a market that now uses server farms instead of a single work station to enable its functions. I do not know if this current generation of chip will be able to replace the server farm, but if not this chip cycle, then the next, and server farms will start to fall to single work stations for those in the movie and real-time graphics industries.

Tinker

11 Likes

Using TMFCochrane’s numbers I am projecting EPS this year of $10.49, leaving a trailing P/E in 6 months of ~24.

JPM is using FY19 estimates of $8.75, an FY20 of $9.10. Price target of $265.

They would be a buyer on pullbacks.

Q3 consensus is $1.99, so your rear-view mirror method of forecasting may be off as you note.

Your estimate is 20.0% higher with the year already halfway over.

Just fyi.

3 Likes

What NVDA also announced is that the new chip set will open up an entirely untapped market in computer graphics that is enormous. It is a market that now uses server farms instead of a single work station to enable its functions. I do not know if this current generation of chip will be able to replace the server farm, but if not this chip cycle, then the next, and server farms will start to fall to single work stations for those in the movie and real-time graphics industries.

In FY2018 the profession visualization revenue for NVDA was $933M. Last quarter it was $281M. CEO Huang said that the visual effects industry is $250B, but how much of that industry is addressable to NVDA’s Turing chips? I think that the ProVis segment that NVDA breaks out in it’s revenue reporting will begin to accelerate in growth starting in Q4. I do think that real-time ray tracing will have a big impact on gaming and on VR.

Chris

4 Likes

In FY2018 the profession visualization revenue for NVDA was $933M. Last quarter it was $281M. CEO Huang said that the visual effects industry is $250B, but how much of that industry is addressable to NVDA’s Turing chips? I think that the ProVis segment that NVDA breaks out in it’s revenue reporting will begin to accelerate in growth starting in Q4. I do think that real-time ray tracing will have a big impact on gaming and on VR.

The biggest movies in the world make extensive use of CGI (either to augment real world shots or as straight up animation). Marvel’s entire franchise has been made possible by advances in rendering, though if you go back and watch Iron Man you still see how much better CGI is now compared to before.

Here is an article describing what Turing is allowing computer animators to do:
https://www.chaosgroup.com/blog/ray-traced-tendering-acceler…

From what I can tell only in the past couple years have GPU enabled render farms become available. And since it requires so much power, smaller film makers and studios have to pay a lot for cloud based rendering. This will potentially make it cost effective for computer artists to build their own rendering machines. Render farms can upgrade to Turing and save on energy costs (the more you buy the more you save). And CGI will become available to many more artists/film makers as the cost to render photorealistic video drops.

Gaming is clearly the next announcement, though it will be a while before games take advantage of the real time ray tracing. I think it will be a couple years before usable for augmented/virtual reality since you need even higher frame rates, but it’s fun to imagine the possibility of augmented reality with real time ray tracing detecting ambient lighting and casting realistic shadows/reflections on simulated objects.

As Chris mentioned, it’s unclear what portion of the $250b total market NVDA could capture, but I’m guessing it’s significantly more than 0.4%.

10 Likes

Cross post from July 1 on NPI. I had stumbled on Chaos Group through a Google Cloud link. Chaos started to ramp use of NVDA GPUs for their Render service and partnered with Google Cloud to scale out rendering on V100 GPUs. They highlight the Black Panther movie and the use of their VRay GPU service to make some of the complex and visually stunning scenes.

http://discussion.fool.com/nvda-and-wakanda8217s-scenery-3311083…

Like with most computing in the world, most of rendering is still being done on racks and racks of CPUs. Nvidia and GPUs and almost a million CUDA developers are disrupting that. This is one example of thousands.

Nvidia’s pro segment group Accelerated growth this last quarter. Went from low teens to mid twenties growth I believe. I still think this will be a sleeper segment for them that will surprise the market this year.

Chaos claimed an 80% increase in render performance with Volta Quadro GPUs and loved it so much they renamed the product from VRay RT to VRay GPU. This was pre Turing, this could be huge. Could level off in Q3 in prep for Turing launch, but Q4 might be really interesting.

Darth

4 Likes

Dang! That is yet another market dominated by CPUs that GPUs will disrupt in the next 3 to 5 years and that is a very large market. The advantages of the GPU in regards to efficiency and economics is far too great for this not to happen. This will also further democratize creative cinema down to the UTube level, much less all the existing commercial businesses who currently now have to use server farms of CPUs.

One vertical at a time it seems.

Q3 revenue of 23% is the same sort of thing that hit Arista. With Nvidia however, it appears to be just a temporary hiatus with crypto-currency disappearing but multiple other markets either greenfielding are reaccelerating again, such as a whole new generation of gaming chips, not to mention the existing chips moving down the ladder (price wise as well) putting even a larger hurt on AMD as their top level GPU will now be underpriced by the better performing NVidia GPUs, and yet AMD will not have any product to compete with Turing.

Q4 should be quite telling and big this year. Last year the last two quarters had 63% higher earnings than the first 2Qs, the year before it was more than 100% higher, and in every year back to 2014 the Q3 and Q4 have materially had greater earnings than the first 2Qs. This year the gap may be less than usual given the crypto-currency reduction that existed in Q1, and Q3 slow down possibly in gaming GPU chips depending on the roll out of Turing. But Q4 ought to be quite nice.

Even if the premium is only 20% this year in the back two quarters (the lowest premium since 2014 as far as I can tell, by far) we are still looking at more then $8.00 eps a share with revenue growth and earnings growth reaccelerating again. However, I have no doubt Nvidia will conservatively guide for Q4 in Q3, and not try to estimate upward for pent up demand.

Excellent find Darth. Turing takes the information you gave and puts it on steroids.

Tinker

3 Likes