Its accounted for in their "GeForce" segment, which pertains specifically to the consumer discrete GPU line.
boxleitnerb
Regular
Do you have any proof of that?
But as I said, even without it, they make plenty of profit in excess of 100m.
NVIDIA's consumer GPU business includes the following: desktop, notebook, memory, chipset products, and licensing revenue from their patent cross-license agreement with Intel. This information is directly from their CFO commentary.
The licensing revenue from Intel is in the form of annual payments (between $200-$300 million per year, $1.58 billion total recognized over a six year term), not quarterly payments. [strike]I believe that this licensing revenue is reflected in Q4 of each fiscal year (not calendar year).[/strike]
The consumer (Geforce) GPU business is clearly very important to NVIDIA, as it provides the economies of scale to pursue higher margin professional (Quadro/Tesla/etc.) solutions. In the future, the consumer GPU business will be bolstered by the Tegra business, and the GPU designs for Tegra will serve as the building block for future discrete Geforce GPU's.
The licensing revenue from Intel is in the form of annual payments (between $200-$300 million per year, $1.58 billion total recognized over a six year term), not quarterly payments. [strike]I believe that this licensing revenue is reflected in Q4 of each fiscal year (not calendar year).[/strike]
The consumer (Geforce) GPU business is clearly very important to NVIDIA, as it provides the economies of scale to pursue higher margin professional (Quadro/Tesla/etc.) solutions. In the future, the consumer GPU business will be bolstered by the Tegra business, and the GPU designs for Tegra will serve as the building block for future discrete Geforce GPU's.
Last edited by a moderator:
Guys please this is getting absurd.
The link given earlier - http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTY5NDk0fENoaWxkSUQ9LTF8VHlwZT0z&t=1 points out that
If Nvidia were getting paid by intel at the end of the year they wouldn't have given guidance between $1.015 and $1.175 billion - it would have been closer to $1.4 billion. They almost certainly get paid in the way Alexko said, that is $75m/qtr.
The link given earlier - http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTY5NDk0fENoaWxkSUQ9LTF8VHlwZT0z&t=1 points out that
This can be seen in Note 14: Segment Information
If Nvidia were getting paid by intel at the end of the year they wouldn't have given guidance between $1.015 and $1.175 billion - it would have been closer to $1.4 billion. They almost certainly get paid in the way Alexko said, that is $75m/qtr.
Last edited by a moderator:
That is incorrect. The licensing revenue from Intel is paid on an annual basis, not a quarterly basis. Also note that NVIDIA does their financial reporting based on fiscal year, not calendar year. For instance, when NVIDIA says Q1 fiscal year 2013, that equates to Feb-Apr calendar year 2012. So the next financial reporting period is Q4 FY 2013.
Note that NVIDIA's revenue in fiscal year Q3 will typically be significantly higher than fiscal year Q4. The reason for this is that products sold by NVIDIA's partners to end users during the holiday season need to be supplied by NVIDIA to it's partners well in advance of the holidays, and therefore most of this holiday season revenue is realized by NVIDIA in fiscal year Q3.
Note that NVIDIA's revenue in fiscal year Q3 will typically be significantly higher than fiscal year Q4. The reason for this is that products sold by NVIDIA's partners to end users during the holiday season need to be supplied by NVIDIA to it's partners well in advance of the holidays, and therefore most of this holiday season revenue is realized by NVIDIA in fiscal year Q3.
Last edited by a moderator:
It is paid on an annual basis, but Nvidia reports it quarterly basis and divides it by 4 for each quarter. It is in their CFO commentary PDF files on their website. But it is irrelevant to the points trying to be made. Nvidia has essentially 3 sources of revenue: money from GPU's, money from SoC's, and money from licensing agreements (namely the PS4). Nvidia lost about $250-300 from their SoC division last year (JHH has said the break even point for Tegra is about $1 billion in annual sales). Take away that $300 million loss, and suddenly nvidia's profits look pretty damn good overall.
You are correct, my apologies, the licensing revenue from Intel is reported on a quarterly basis (although paid on an annual basis). That makes sense too, as there is a smoother trend line from one quarter to another. It appears that licensing revenue from Intel is reported as ~ $66 million per quarter and will remain flat at that level until the six year duration is completed (for a grand total of $1.58 billion reported over 24 quarters in total).
Why? There's nothing conflicting about PC's using discrete GPU's while consoles use APU's.
PC's can potentially even use APU's for GPGPU alongside discrete GPU's for graphics.
The point is that PCs aren't going to need discrete graphics to keep up with consoles if consoles themselves don't have it.
Of course, nothing says PCs can't be faster than consoles, but the latter are not going to fuel discrete GPU development.
I see your point,m but the same could have been said about the current generation once we hit G80 levels of performance. There was no drive from the consoles for additional performance - but we got it anyway.
Yes, we got it, but not because of consoles.
DavidGraham
Veteran
That point is moot anyway , consoles will NOT have APUs .
How do you know that? It's only rumors at this point.
I wonder from where exactly that certainty comes from for the majority of upcoming consoles. Better can I use that phrase in my signature to have something to laugh at in about 2 weeks?
The whole discussion about GeForce profitability blah blah is pointless if the following question isn't answered: would the professional business make as much profit as it does if there was no GeForce business.
We all know the answer to that, I should hope: it'd be saddled with a mountain of NRE and would bleed red ink despite high margins. Ready to be wiped away by a competitor who can leverage consumer designs. (See Silicon Graphics, 3D Labs etc.)
IOW: looking at the GeForce business on its own an exercise in futility.
Look at the GPU business combined and you have a much better idea about whether its worth participating in it. Nothing that hasn't been said in the past... Ad nauseam.
Feel free to try make the same case for AMD leveraging its GPUs into APUs. (With the margins they have on CPUs and Intel giving its GPU away for free, I have a hard time seeing a parallel, but who knows...)
We all know the answer to that, I should hope: it'd be saddled with a mountain of NRE and would bleed red ink despite high margins. Ready to be wiped away by a competitor who can leverage consumer designs. (See Silicon Graphics, 3D Labs etc.)
IOW: looking at the GeForce business on its own an exercise in futility.
Look at the GPU business combined and you have a much better idea about whether its worth participating in it. Nothing that hasn't been said in the past... Ad nauseam.
Feel free to try make the same case for AMD leveraging its GPUs into APUs. (With the margins they have on CPUs and Intel giving its GPU away for free, I have a hard time seeing a parallel, but who knows...)
DavidGraham
Veteran
The Wi U didn't have it , the leaked specs of the upcoming Xbox720 and PS4 don't mention anything about APUs , rather the opposite actually , large and powerful discrete GPUs are included .
boxleitnerb
Regular
But APUs as well, right? Wouldn't be a bad idea - use the APU for physics which should work well considering the low latency between CPU and GPU part and the discrete GPU for graphics.
DuckThor Evil
Legend
The GPU and CPU are in the same die in PS4 and Durango.
DavidGraham
Veteran
Hmmm , really? I didn't know that. Sorry.. my mistake ,I guess that's why people are calling it an APU then .
However I still think it is really just semantics . What is the difference between a powerful discrete GPU and a powerful integrated APU ? their place relative to the CPU ? well that's not really an earth shattering difference.
The fact of the matter is , so long as consoles use decent or powerful GPUs (intergratd or not), PCs will need decent discrete GPUs to match them and to exceed .. we all know PCs lag behind in the optimization and utilization front , they will need to compensate for that with more powerful hardware .
UniversalTruth
Veteran
No one will ever use this term "powerful integrated APU". First of all because the nature of the APU itself already determines that it would suck very badly in comparison to normal discrete CPU+ GPU. There are too many bottlenecks appearing when you put them on a single die.
You would probably mean "powerful integrated GPU in an APU".
That is true. The industry is evil enough to unleash games which run artificially very badly on existing hardware in order to push sales...
