NVIDIA Kepler speculation thread

[A1xLLcqAgt0qc2RyMz0y]

Not really. The PC business is stagnating faster than their Tegra market share is rising.

The actual numbers don't back up your assertion:

GAAP Annual Revenue Comparison (in millions)

FY13 FY12 Y/Y$ Y/Y%
GPU $3,251.7 $3,186.8 up $64.9 up 2.0%
Tegra $764.5 $591.1 up $173.4 up 29.3%

CFO Commentary on Fourth Quarter 2013 Results

http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTcxNjMzfENoaWxkSUQ9LTF8VHlwZT0z&t=1

 

[A1xLLcqAgt0qc2RyMz0y]

Indeed. (and have they actually said they're making custom CPU cores?

Yes, Project Denver.

Nvidia: Denver Is Custom ARMv8 64-Bit Processor with Secret Sauce.

At present, the only thing clear about Nvidia Denver is that it is not based on pure Cortex-A50 cores, but on custom Nvidia-developed cores which are presumably more advanced. In addition, Mr. Huang reemphasized that Denver will be also positioned for servers, though he admitted that it would take some time before ARM will get popular in servers in general. Interestingly, but the head of Nvidia did not mention another interesting project that is developed by Nvidia, project Boulder, which is projected to take on AMD Opteron and Intel Xeon high-performance server chips.

“We are big believers in ARM-based servers. […] We have been working on project Denver for two years now, so we are going to have custom ARM designs for this market,” said Mr. Huang in a conversation with the Tech Trader Daily blog.

http://www.xbitlabs.com/news/cpu/di...ARMv8_64_Bit_Processor_with_Secret_Sauce.html

 

[A1xLLcqAgt0qc2RyMz0y]

Revenue was well down on Q3 but they still had a decent quarter overall.

From CFO commentary: Revenue for the fourth quarter of fiscal 2013 was $1.11 billion, down 8.1 percent sequentially and up 16.1 percent year-over-year. Our outlook was $1.025 billion to $1.175 billion.

So the revenue was dead center from what they had predicted a quarter ago and was up 16.1% Y-on-Y.

Their outlook for Q1 is way down on what the market thought it would be though so there must be a pretty high chance of the stock going south today.

The revenue reduction was explained in the conference call (current PC slowness (Q1 is usually slow anyway) and transition from T3 to T4)

http://seekingalpha.com/article/118...results-earnings-call-transcript?source=yahoo

As for the stock going south today it is currently up $0.21 (1.7%).

 

[jimbo75]

So the revenue was dead center from what they had predicted a quarter ago and was up 16.1% Y-on-Y.

The revenue reduction was explained in the conference call (current PC slowness (Q1 is usually slow anyway) and transition from T3 to T4).

Sure Q1 is slow but it's the size of the projected dip that has analysts worried. It's almost $70m less than what they were predicting - it's not a typical slow Q1 same as Q4 wasn't "seasonally down". This is here to stay. Nvidia is actually doing very well with a pretty poor hand tbh, but Tegra must be a huge worry for them.

 

[jimbo75]

The actual numbers don't back up your assertion:

CFO Commentary on Fourth Quarter 2013 Results

http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTcxNjMzfENoaWxkSUQ9LTF8VHlwZT0z&t=1

The problem with that is their operating expenses have gone up even more and net income is down YoY. Tegra is dragging the company down, but they know it's the only future they have so they have to stick with it. With Tegra 4 being undesirable the short term looks pretty bad.

 

[ams]

Q4 revenue is not typically lower than Q3 revenue. In 2011 it was lower, in 2010 and 2009 it was higher.

The sensible comparison is Q4/Q3 Fiscal Year 2013 vs. Q4/Q3 Fiscal Year 2012. The fiscal years prior to FY 2012 had a very different product mix, with significant revenues from the chipset business and insignificant revenues from the Tegra business.

How long ago was it they started investing in Tegra? Since then it's just been a money sink and they aren't really close to breaking even. Their OpEx has ballooned as they've hired more and more people into an unprofitable business.

Investment in Tegra is a multi-faceted and ongoing effort, but there is no question that Tegra 4 (Wayne), Tegra 5 (Logan), and Tegra 6 (Stark) are all fully funded. NVIDIA's investment in software defined baseband modem technology started less than two years ago. NVIDIA's investment in custom low power CPU's started more than five years ago. Designing a fully custom CPU takes time and money, and there is no way to get around that. Tegra 4 is an admirable effort from NVIDIA and will unquestionably be one of the highest performance SoC's available in 2013, but NVIDIA will not fully hit their stride until Tegra 5, which should incorporate a fully custom CPU, a fully custom GPU with unified CUDA "cores", and the latest-and-greatest 4G LTE software defined baseband processor.

 

[ams]

This presupposes that others investing in those same areas will be less successful than Nvidia.

Not really, because the TAM (total addressable market) for smartphones, tablets, 4G LTE equipped devices, Android gaming devices, cloud gaming devices, etc. in general is rapidly increasing and will continue to do so over the next few years.

 

[AlphaWolf]

Not really, because the TAM (total addressable market) for smartphones, tablets, 4G LTE equipped devices, Android gaming devices, cloud gaming devices, etc. in general is rapidly increasing and will continue to do so over the next few years.

Maintaining market share even in the projected growing market isn't a path to profitability for Tegra. They need to take something away from Samsung or Qualcomm or another and hope Intel and others are not successful at that same plan. There might be a pot of gold at the end but there are a number of not insignificant obstacles in their way.

 

[jimbo75]

This is Nvidia's problem in a nutshell -

Half of their operating income is royalty payment from Intel. Much of the other half is under threat from Intel either through improved graphics in Haswell, or Xeon Phi - all the while Tegra loses money.

It's not like they are going under any time soon (mainly thanks to a large bank balance) but they are in a weak position.

 

[ams]

Maintaining market share even in the projected growing market isn't a path to profitability for Tegra.

The goal is obviously to increase market share in a growing market. To increase market share, NVIDIA will need to take away share from competitors such as Qualcomm, Samsung, etc. But even if market share is maintained, revenues (and profitability) will tend to improve if the TAM improves.

 

[AlphaWolf]

The goal is obviously to increase market share in a growing market. To increase market share, NVIDIA will need to take away share from competitors such as Qualcomm, Samsung, etc. But even if market share is maintained, revenues (and profitability) will tend to improve if the TAM improves.

Not really because the costs are likely to go up.

 

[tviceman]

This is Nvidia's problem in a nutshell -

Half of their operating income is royalty payment from Intel. Much of the other half is under threat from Intel either through improved graphics in Haswell, or Xeon Phi - all the while Tegra loses money.

It's not like they are going under any time soon (mainly thanks to a large bank balance) but they are in a weak position.

The company just started a quarterly dividend Q4 last year. They bought $100 million dollars of their own shares during the last quarter. Do you think a company that is struggling to survive long term does either of those things? Let me answer that for you - NO.

Tegra 4 is coming in April-May, that will drive sales back up for them. They are sampling their own LTE modem. That will be a new area of growth for them. Tegra 5 will give them access to the server market - that will be another new area of growth for them. Tegra grey is coming Q3/Q4. That will be a new area of growth for them. They are in a weak position for the current quarter and current quarter only. They are investing heavily to diversify their product lineup, and are gaining momentum. Tegra isn't profitable yet but it's getting there. Tegra is eating more money than is being made up by the Intel settlement, but use graphs and projections and it's not hard to see that Tegra will likely be profitable before Intel's settlement payments end.

Titan is coming next week, which is mostly bragging rights but nevertheless it will spawn refreshes and new products to replace the existing GPU lineup. Quadro cards are launching en masse right now. Next year they'll be able to offer 100% independent server solutions (no Intel / AMD chips needed). They'll be able to ffer 100% independent HPC solutions. They'll have gone from zero LTE modem sales to XXX amount of LTE modem sales.

 

[ams]

Half of their operating income is royalty payment from Intel.

If you look at Q4 FY 2013, the licensing revenue from Intel represents about 6% of NVIDIA's total GAAP quarterly revenue, and about 30% of NVIDIA's total GAAP net income. So the licensing revenue is significant, but not the panacea it is made out to be. Intel antics did cause serious harm to NVIDIA's chipset business, and this licensing revenue [more than] replaces the revenue lost from NVIDIA's exit out of the chipset business.

 

[ams]

Not really because the costs are likely to go up.

R&D costs are fixed costs. Fabrication of extra quantities (in order to meet extra demand) is a variable cost. As long as products are sold at a profit and not a loss, then increases in TAM will result in increases in profit, provided that market share is maintained.

 

[Ailuros]

The actual numbers don't back up your assertion

Now skip the revenue values and replace them volumes. Or are you trying to tell me that Tegra has any worth mentioning market share in the SFF mobile market at the moment and especially in smartphones?

 

[AlphaWolf]

R&D costs are fixed costs. Fabrication of extra quantities (in order to meet extra demand) is a variable cost. As long as products are sold at a profit and not a loss, then increases in TAM will result in increases in profit, provided that market share is maintained.

A larger market should stimulate greater investment from competitors. And costs are ever increasing regardless.

 

[Alexko]

Now skip the revenue values and replace them volumes. Or are you trying to tell me that Tegra has any worth mentioning market share in the SFF mobile market at the moment and especially in smartphones?

Grey may help (if its modem is actually competitive, which remains to be seen) but it doesn't eliminate one big problem:

Much of the market is Apple + Samsung. Apple uses its own SoCs exclusively, and Samsung usually does too. Nokia is trending down, as is HTC. As far as I'm aware RIM doesn't use high-end SoCs, so NVIDIA's out.

That leaves a myriad of small manufacturers, and of course competition from the much bigger Qualcomm and others, Intel included. With Temash and especially its successors, AMD too is making a play for the tablet market. It's going to be really tough for NVIDIA, and I'm not sure that having its own CPU core design is more of an advantage than a burden for such a small company. If it turns out to be better than Cortex A-57, that's a small advantage for a significant investment. If it's not, it's just dead weight.

 

[ams]

Much of the market is Apple + Samsung.

Yes, Apple and Samsung do have very significant marketshare in the smartphone space, but the TAM outside of Apple and Samsung is clearly trending upward too.

It's going to be really tough for NVIDIA, and I'm not sure that having its own CPU core design is more of an advantage than a burden for such a small company. If it turns out to be better than Cortex A-57, that's a small advantage for a significant investment. If it's not, it's just dead weight.

NVIDIA employs literally thousands of engineers. Considering that they are able to design and engineer some of the fastest and most complicated GPU's in the world, and considering that they have hired a lot of extra talent to work on CPU's, I don't think it would be above them to design a world-class CPU. We'll just have to wait and see. Note that the end goal here is not just to design a good CPU, but to integrate a fully custom NVIDIA CPU on the same chip as a fully custom NVIDIA GPU.

 

[Silent_Buddha]

R&D costs are fixed costs. Fabrication of extra quantities (in order to meet extra demand) is a variable cost. As long as products are sold at a profit and not a loss, then increases in TAM will result in increases in profit, provided that market share is maintained.

That's generally true in a simplified manner but not absolutely true. This assumes that profit margins remain healthy.

In a case where operating costs increase, but profit margins decrease, you could end up with a situation where the product itself is profitable, but once operating costs for that division are factored in, you end up with a net loss.

BTW - this isn't me disagreeing with you that it is possible for Nvidia to achieve profitability while maintaining stable market share (neither losing nor gaining) as long as TAM increases. But there's no guarantee that it will happen.

And while R&D is a fixed cost per generation of products, if you are to maintain healthy profit margins and thus achieve profitability, you must always have ongoing R&D in order to regularly have new products. Hence R&D can also be considered a variable cost due to changing market conditions and requirements for competitive products (IE - R&D can increase or decrease depending on the level of competition and requirements for competitive products).

Regards,
SB

 

[A1xLLcqAgt0qc2RyMz0y]

This is Nvidia's problem in a nutshell -

Half of their operating income is royalty payment from Intel.

And that is a problem exactly why?

Intel is basicly paying Nvidia a royalty stream for the right to use Nvidia's patents in their internal GPUs in processors and for Xeon Phi.

For someone who constantly posts how wonderful the royalties AMD will get from the consoles Intel's royalty payments to Nvidia are constant and pure profit to the bottom line.

AMD got a one time $1.25 billion from Intel and after the initial boost has been back posting quarter after quarter losses.

Nvidia got $1.5 billion spread out over six years and even if you back it out still makes a profit.

Oh, and when the cross licensing agreement comes up for renewal expect it to be renewed and the revenue stream to continue.