Nvidia to license Kepler to everyone who can pay

[blackened23]

Sorry, but what? Their Tegra business grew last year.

Let me guess you have no clue and doing viral marketing for a certain company, right?

LOL...Yeah, that's obviously it W.R.T. your reference to siliconwars. The guy with 40,000 references to Tegra and nvidia to his nickname on google search isn't though. I mean, it's just a hobby to spend that much time talking Tegra and nvidia up, yeah? Anyway, we've discussed this - Tegra has yet to actually make a profit for nvidia. Sure, sales are growing but the point remains: He stated that Tegra did not make a net profit for nvidia yet. That was a correct statement. If you look at the divisional breakdowns, the Tegra division has more expenditures than revenue. Thus it has not been profitable yet - we've already discussed this many times.

That's not necessarily a problem, yet - as someone else mentioned, a company has to spend money to make money in the end. It's all part of the growing pains. What's particularly worrisome in this case, however, is that nvidia lost some major design wins this year including the MS Surface RT (they switched the Qualcomm) and the Nexus 7/10 refresh. (also using Qualcomm). Looking at the benchmarks for the new Qualcomm part, it's easy to see why a lot of vendors are using it - I've even seen some news indicating that Asus will be using the new Qualcomm chip in their upcoming 2013 tablets. Is that a problem for nvidia? I don't know, you tell me. Now I have heard that Toshiba and HP are using the Tegra 4 for some upcoming tablets. I guess we'll see what happens, although the delay for Tegra 4 probably hurt it. I can see valid arguments being made for NV licensing their technology in response to the Tegra 4 delays and the losses of key design wins, although it is complete speculation. None of us know the inner workings over at NV, so it's anyone's guess as to why they're doing this. I will say one thing though, the ARM SOC market seems to be pretty "dog eat dog" in terms of competition, the competition is pretty fierce. I can't see the 1 quarter delay of Tegra 4 being beneficial at all, especially when the new Qualcomm part is shipping now (IIRC).

2013 Surface RT using Qualcomm Snapdragon 800 :

http://www.theverge.com/2013/6/19/4444198/surface-rt-qualcomm-processor

 

[SiliconWars]

Sorry, but what? Their Tegra business grew last year.

Growing with growing losses doesn't really impress me much.

Let me guess you have no clue and doing viral marketing for a certain company, right?

Qualcomm doesn't need any viral marketing, they have their chips in all the relevant products already.

 

[sontin]

LOL...Yeah, that's obviously it W.R.T. your reference to siliconwars. The guy with 40,000 references to Tegra and nvidia to his nickname on google search isn't though. I mean, it's just a hobby to spend that much time talking Tegra and nvidia up, yeah?

Sorry, that i'm writing about nVidia here. Maybe someone should put out a notice that nVidia is now a banned word. :\

Anyway, we've discussed this - Tegra has yet to actually make a profit for nvidia. Sure, sales are growing but the point remains: He stated that Tegra did not make a net profit for nvidia yet. That was a correct statement. If you look at the divisional breakdowns, the Tegra division has more expenditures than revenue. Thus it has not been profitable yet - we've already discussed this many times.

Tegra is making money because they are selling every chip with a ~50% gross margins. If they would stop to invest then they would have a positiv operate income.

That's not necessarily a problem, yet - as someone else mentioned, a company has to spend money to make money in the end. It's all part of the growing pains. What's particularly worrisome in this case, however, is that nvidia lost some major design wins this year including the MS Surface RT (they switched the Qualcomm) and the Nexus 7/10 refresh. (also using Qualcomm).

It's only certain that they lost N7. Everything else are rumours like that nVidia still got one Surface win...

Looking at the benchmarks for the new Qualcomm part, it's easy to see why a lot of vendors are using it - I've even seen some news indicating that Asus will be using the new Qualcomm chip in their upcoming 2013 tablets. Is that a problem for nvidia? I don't know, you tell me. Now I have heard that Toshiba and HP are using the Tegra 4 for some upcoming tablets. I guess we'll see what happens, although the delay for Tegra 4 probably hurt it. I can see valid arguments being made for NV licensing their technology in response to the Tegra 4 delays and the losses of key design wins, although it is complete speculation. None of us know the inner workings over at NV, so it's anyone's guess as to why they're doing this. I will say one thing though, the ARM SOC market seems to be pretty "dog eat dog" in terms of competition, the competition is pretty fierce. I can't see the 1 quarter delay of Tegra 4 being beneficial at all, especially when the new Qualcomm part is shipping now (IIRC).

Asus will use Tegra, CloverTrail+ and maybe Snapdragon. Looks alright for me.
And Tegra 3, Omap4 and Krait were supported plattforms by RT. There was always competition.

 

[tviceman]

Ask yourself this more easily answered question: if no one had already expressed interest, or if Nvidia does not think they will gain any licenses out of this move, why would they announce such a thing in first place?

Can you answer that question for me, because I have no idea.

The obvious answer is they wouldn't announce such a move if there is no market for it.

And that still does not answer the question, who will license Kepler?

Potentially anyone who designs and makes a system-on-a-chip processor? Or is that too far fetched? http://semiaccurate.com/2013/06/19/nvidias-kepler-license-has-no-chance-of-success/ I mean, obviously, this guy has been so right, so many times about Nvidia's financial distress, failures at manufacturing, market share, leadership, etc. that he is obviously spot on once again! Didn't we all realize, Nvidia actually went out of business 4 years ago. They never once made a single dollar nor did they ever have a clue as to what they were doing building GPU's or running their company.

 

[ams23]

This is not rocket science here folks. NVIDIA's business model will continue as is for the majority of areas that they currently serve (including Tegra/Geforce/Quadro/Tesla). What NVIDIA is trying to do is to go after business that they would not have otherwise gotten due to lack of flexibility in meeting the needs of large vertically integrated clients. At the end of the day, NVIDIA's revenue from selling their own graphics processing units will dwarf their revenue from licensing graphics technology for many years to come.

Tegra is a nascent business, and costs tend to outweigh revenues when a nascent business is built up. That said, I do believe that NVIDIA is finally at or near the point where Tegra is actually a profitable business for them, because the incremental investment needed for Tegra moving forward is relatively small compared to their overall investment in graphics and visual computing technology. Starting with Tegra 5, the ultra low power GPU's used in Tegra will feature a modern GPU architecture from the company, and NVIDIA will be able to heavily leverage their investments in the GPU.

In order for Tegra to gain some traction in the ultra mobile smartphone space, NVIDIA really needs to have their own baseband modem processor that is certified on different networks. I'm sure they are working very hard on doing that. In the meantime, Tegra has had some success in the tablet space, and quite a bit of success in the automotive space (with $2+ billion USD of design wins lined up for the next few years). In fact, revenues for Tegra automotive on it's own are projected to be over $450 million USD by FY2016, if not even higher at some point in the future.

 

[Siberian]

Tegra is making money unlike AMD. NVIDIA is the graphics leader by far in terms of marketshare and performance. Looks like a smart move to get money from comapnies that make their own SoC.

Some people are so filled with hate it clouds their judgement. :thumbsdown:

 

[Stuka87]

Tegra is making money unlike AMD. NVIDIA is the graphics leader by far in terms of marketshare and performance. Looks like a smart move to get money from comapnies that make their own SoC.

Some people are so filled with hate it clouds their judgement. :thumbsdown:

Tegra is not making money. The whole devision has not had a positive quarter, ever (5 seconds on google will show plenty of reports on this).

 

[joshhedge]

Tegra is making money unlike AMD. NVIDIA is the graphics leader by far in terms of marketshare and performance. Looks like a smart move to get money from comapnies that make their own SoC.

Some people are so filled with hate it clouds their judgement. :thumbsdown:

Certainly not a graphics leader in the SoC market, Tegra is weak.

 

[Tsavo]

Given how goofy AMD are, I wonder how long it will be before we can get an AMD GTX Radeon.

 

[blastingcap]

Nvidia stock prices rose today on the news, so the market seems to think it's a profitable idea.

 

[SirPauly]

Let me refresh your memory:

- In 2009 Intel dropped FSB for Nehalen and 2010 with Westmere, effectively shutting Nvidia out of their lucrative chipset business.

- Nvidia sued Intel for their IGPU IP, while Intel sued Nvidia to not allow them to develop chipset products for their brand new processors.

- Nvidia complained to the FTC, which was behind Intel at the time and had lost AMD once they settled with Intel.

- The FTC then started to build a case against Intel, with Nvidia collaborating.

- Intel settled with the FTC and Nvidia by paying Nvidia 1.5 billion in five years *and* not giving them the much coveted x86 license.

So I don't think it's fair to say that this is money out of a license agreement. This was an agreement done with the FTC holding a gun on Intel's head, it would be pay or go through a lenghty judicial fight with unpredictable outcome.

My point was simple:

nVidia already has a licensing agreement with Intel. Don't know what you're arguing against!

 

[Siberian]

Tegra is not making money. The whole devision has not had a positive quarter, ever (5 seconds on google will show plenty of reports on this).

http://www.bit-tech.net/news/hardware/2013/02/14/nvidia-record-profits/

Record profits driven by Tegra growth.

 

[Rezist]

http://www.bit-tech.net/news/hardware/2013/02/14/nvidia-record-profits/

Record profits driven by Tegra growth.

They never mention where the profits come from however. They mention Tegra revenues but not there profit margins by division

 

[hawtdawg]

Does this mean that Intel could make their own Nvidia GPU's using their own fab if they wanted?

 

[tviceman]

Does this mean that Intel could make their own Nvidia GPU's using their own fab if they wanted?

Hahahaha wow that would be a kick in the nuts if they went with their most advanced fabs!

 

[Keysplayr]

They never mention where the profits come from however. They mention Tegra revenues but not there profit margins by division

so what does that mean? To you?

 

[blackened23]

It's pretty simple to anyone who can properly interpret fiscal results. The article is in error, they looked at the total profit of nvidia as a whole and not the divisional breakdowns. Of course nvidia overall made a profit based on their geforce and professional HPC sales. They do have statements with the Geforce, Tegra, and professional business segments separated, and the tegra division has expenditures exceeding revenue and has not turned a profit yet. Yet, it is growing. Yes, this is part of spending money to make money, the initial investment is the biggest opportunity cost. But Tegra has not turned a net profit yet.

Also, 60% of nvidia's Tegra 3 sales were with the Google Nexus. That will no longer be using nvidia Tegra. Perhaps nvidia can pull a rabbit out of the hat, but I don't see Shield being that rabbit, I personally feel shield is doomed to fail or will be an extremely low volume niche product.

 

[Siberian]

It's pretty simple to anyone who can properly interpret fiscal results. The article is in error, they looked at the total profit of nvidia as a whole and not the divisional breakdowns. Of course nvidia overall made a profit based on their geforce and professional HPC sales. They do have statements with the Geforce, Tegra, and professional business segments separated, and the tegra division has expenditures exceeding revenue and has not turned a profit yet. Yet, it is growing. Yes, this is part of spending money to make money, the initial investment is the biggest opportunity cost. But Tegra has not turned a net profit yet.

Also, 60% of nvidia's Tegra 3 sales were with the Google Nexus. That will no longer be using nvidia Tegra. Perhaps nvidia can pull a rabbit out of the hat, but I don't see Shield being that rabbit, I personally feel shield is doomed to fail or will be an extremely low volume niche product.

Feel free to post links to back up what you say.

 

[ams23]

Tegra should easily be a profitable business for NVIDIA moving forward. IIRC, the incremental investment for Tegra during [Fiscal Year] FY2014 will be "only" ~ $300 million USD, whereas revenues for the overall Tegra business (including tablets, convertibles, smartphones, automotive, embedded) will be at least 2x higher than this. So even after allocating expenses, the Tegra business will still be profitable this fiscal year and in the future.

 

[Rezist]

so what does that mean? To you?

That the stated link does nothing to disprove what's been discussed??? Which is the profitability of the Tegra division. That probably didn't need explaining.
Anyways ARM will most likely be the ones to incorporate these GPU's since they are the farthest behind are they not?

 

[beginner99]

Anyways ARM will most likely be the ones to incorporate these GPU's since they are the farthest behind are they not?

Was my thought too. Maybe they even "cross-license". NV gets ARMs CPU and they in return NVs GPU IP.

 

[IntelUser2000]

Does this mean that Intel could make their own Nvidia GPU's using their own fab if they wanted?

They'd never do that though. It'd be an extremely low volume business with lower margins than making their own.

Anyway, it isn't total loss for Nvidia for Surface. The next Surface RT is supposed to have two variants at least, one with Qualcomm, and another with Nvidia.

 

[beno619]

Bah Tegra 3 isn't horrific… It's just not great. You've just got to know which custom kernels and ROMs to use on the One X to unlock some decent performance and improved battery life.

I'm running ARHD and have played around with custom kernels, back to the stock one right now.

Doesn't take away from the fact Tegra 3 is a joke.

You would have to be crazy if you believe it's a good soc.

 

[Imouto]

Feel free to post links to back up what you say.

Did you feel any good posting the only article with wrong info in the net?

http://phx.corporate-ir.net/phoenix.zhtml?c=116466&p=irol-SECText&TEXT=aHR0cDovL2FwaS50ZW5rd2l6YXJkLmNvbS9maWxpbmcueG1sP2lwYWdlPTg3OTI1NTkmRFNFUT0wJlNFUT0wJlNRREVTQz1TRUNUSU9OX0VOVElSRSZzdWJzaWQ9NTc%3d

Get to the Note 17 - Segment Information



That's the official earnings breakdown. There you have the numbers for the last 3 years. Now, if you please, stop ashaming yourself.

 

[NTMBK]

Ask yourself this more easily answered question: if no one had already expressed interest, or if Nvidia does not think they will gain any licenses out of this move, why would they announce such a thing in first place?

To prop up the share price a bit longer, amid the bad publicity from not winning a single console or announcing any Tegra 4 phones. And if they did have someone interested in licensing it, they would have announced that. This is closer to their Tegra 4 press conference where they had no design wins, so they had to bring out their own handheld (with a fan).