NVIDIA shows signs ... [2008 - 2017]

[Blazkowicz]

Well, it's not really unheard of that a tech company licenses technology from an another company. Given that AMD is currently somewhat lacking the focus as they are trying to work on multiple fronts (x86 desktops & servers, ARM servers, GPUs), who knows, maybe someday they decide to take the Intel's route and to acquire their GPU technology from an another company and to invest their own research resources somewhere else.

Intel acquired GPU technology in the 90s from Lockheed Martin, that's quite a while ago, a bit less long ago ATI acquired ArtX too which gave them the R300 and the Xbox 360 (and nvidia used the 3dfx Rampage's 2D engine in the geforce FX rather than to make an awesome glide compliant and supersampling monster GPU, but I digress)

Still, I don't think there's really i740/i810 tech in the *bridge or *well GPUs, or Game Cube tech in the Radeon 7970.

 

[3dcgi]

http://phx.corporate-ir.net/phoenix.zhtml?c=116466&p=irol-sec


So they are pretty much admitting they cannot or will not be able to compete with Intel/AMD in the APU front and since Tegra is hardly getting any traction, they are quickly moving to Plan B.
From all this I get the clear impression that Tegra might be on it's deathbed the final blow delivered by Qualcomm's Snapdragon 800 . Who is going to want to license Nvidia GPU IP when such licensee will have to compete with Nvidia itself as they are still producing their own SOCs? It's the same thing Microsoft did to OEMs with Surface. "You can manufacture Surface like tablets but you'll have to pay us for the OS and we'll just come and undercut you by building our own and saving the OS license fee." Remind me how that turned out for Microsoft.

Following that logic who would want to create their own ARM compatible core as they're competing with ARM's design or why would Samsung want to license IMG's designs since Apple is already using them? We know companies do both of these things so the situation isn't as black and white as you describe.

 

[Blazkowicz]

Another thought, AMD going into ARM servers is precisely what you wish for, it's about using other people's tech (straight ARMv8 from ARM, the magical SeaMicro fabric they've acquired, various controllers and stuff that get in the SoC).
Or they included some "ARM trustzone" in some APUs.

 

[3dcgi]

Intel acquired GPU technology in the 90s from Lockheed Martin, that's quite a while ago, a bit less long ago ATI acquired ArtX too which gave them the R300 and the Xbox 360 (and nvidia used the 3dfx Rampage's 2D engine in the geforce FX rather than to make an awesome glide compliant and supersampling monster GPU, but I digress)

Still, I don't think there's really i740/i810 tech in the *bridge or *well GPUs, or Game Cube tech in the Radeon 7970.

The ArtX team didn't work on the Xbox 360 though your point stands. Companies will integrate new technologies or teams, but rarely will they completely license core technology from a main competitor.

Edit: I should add that I mean core technology they already have. Licensing ARM CPU cores fits as a technology AMD didn't have.

 

[Blazkowicz]

AMD licensing the Cortex A57 muddies things, Intel licensed PowerVR GPU technology too.
AMD historically made CPUs, it can be astonishing to see them do this after doing high performance low power CPUs of their own (Geode NX, Athlon Neo, Bobcat, Jaguar).

Their GPU technology can really be seen as a core technology though, one they've bet the company's future on.

 

[jimbo75]

Shield gets a price cut before it's even released -

http://blogs.nvidia.com/blog/2013/06/20/shield-gets-final-touch-before-its-release-a-better-price/

 

[Wynix]

Shield gets a price cut before it's even released -

http://blogs.nvidia.com/blog/2013/06/20/shield-gets-final-touch-before-its-release-a-better-price/

Obviously other factors contributed to the decision, but lowering the price shortly after E3 does seem to lend weight to the idea that the low PS4 price forced their hand.

 

[silent_guy]

So they are pretty much admitting they cannot or will not be able to compete with Intel/AMD in the APU front and since Tegra is hardly getting any traction, they are quickly moving to Plan B.

You missed the part where their business may suffer in case Godzilla tramples headquarters.

 

[silent_guy]

Nvidia's current business can't be taken for granted, but I don't get the negativity about their efforts to expand into new markets.

"They only have a handful of car brands!" Well, yes, but AFAIK this is a market that's only starting to take off. I've seen the insides of a Tesla S and their display is nothing like I've seen in any car before (or since.) There are no guarantees that Nvidia will win in this market, but at least they are there at the start of it all. That's not a bad place to be. ($2B is not bad no matter how you spin it, though I can't find a link to that number?)

"Licensing is a sign of defeat." PR speak or not, when they state that there are vertical markets where they'll never be able to sell real silicon, they have a point. Why shouldn't they try to go after markets that Imagination Tech is currently dominating? Are they supposed to just sit and watch others light up the pixels they have the god given right to drive? It may never replace by their current business completely, of course, but neither does the $60M they get from Intel every quarter, yet it's a nice little extra cash to spend on R&D.

"Shield is a mistake, DOA, etc." According to their investors presentation, it cost $10M to develop (and counting, I assume.) Yeah, it's a gamble. But so what? They'll lose some face here or there if it fails. But probably also learned a great deal in the process.

I thought Grid was a dud (and I'm still unconvinced), but Sony is going to be spending a lot of money on PS3 game streaming to the PS4. If 5 years from now, we look back thinking 'how we didn't see it coming', Nvidia is in a good position now. And Grid for professional applications is something that has a real chance.

They are trying a lot of things. Even if some or most things don't work out, a single hit can do wonders. (Did anyone here believe that putting a discrete Nvidia GPU next to Intel iGPU would go anywhere? It won't make money forever, but it's probably a very large chunk of the notebook GPUs they are selling now.)

 

[Alexko]

You missed the part where their business may suffer in case Godzilla tramples headquarters.

That's only for the Japanese branch.

Nvidia's current business can't be taken for granted, but I don't get the negativity about their efforts to expand into new markets.

"They only have a handful of car brands!" Well, yes, but AFAIK this is a market that's only starting to take off. I've seen the insides of a Tesla S and their display is nothing like I've seen in any car before (or since.) There are no guarantees that Nvidia will win in this market, but at least they are there at the start of it all. That's not a bad place to be. ($2B is not bad no matter how you spin it, though I can't find a link to that number?)

"Licensing is a sign of defeat." PR speak or not, when they state that there are vertical markets where they'll never be able to sell real silicon, they have a point. Why shouldn't they try to go after markets that Imagination Tech is currently dominating? Are they supposed to just sit and watch others light up the pixels they have the god given right to drive? It may never replace by their current business completely, of course, but neither does the $60M they get from Intel every quarter, yet it's a nice little extra cash to spend on R&D.

"Shield is a mistake, DOA, etc." According to their investors presentation, it cost $10M to develop (and counting, I assume.) Yeah, it's a gamble. But so what? They'll lose some face here or there if it fails. But probably also learned a great deal in the process.

I thought Grid was a dud (and I'm still unconvinced), but Sony is going to be spending a lot of money on PS3 game streaming to the PS4. If 5 years from now, we look back thinking 'how we didn't see it coming', Nvidia is in a good position now. And Grid for professional applications is something that has a real chance.

They are trying a lot of things. Even if some or most things don't work out, a single hit can do wonders. (Did anyone here believe that putting a discrete Nvidia GPU next to Intel iGPU would go anywhere? It won't make money forever, but it's probably a very large chunk of the notebook GPUs they are selling now.)

The problem with Shield is that it puts NVIDIA in a position where they're competing with device manufacturers, i.e. with their customers. This might seriously backfire and hurt their Tegra business. Or not. I honestly don't know how manufacturers will react, but NVIDIA could lose more than some face and $10M.

Licensing also presents some potential conflicts of interest, but at worst, it will just make it difficult for NVIDIA to license IP, not hurt their existing business, so I don't see the harm in giving it a shot.

 

[jimbo75]

The way it's meant to be delayed.

 

[jimbo75]

Nvidia Earnings for 2Q 2013 - GPU Strong, Tegra Murdered

 

[ams]

It is unfortunate that Tegra will be well off the pace this fiscal year vs. last fiscal year. Tegra revenue for the entire fiscal year will likely be close to ~ $500 million across all segments (auto, embedded, tablets, smartphones, etc). The saving grace for Tegra is that Kepler.M is already sampling, and that headway is being made in bringing an efficient integrated baseband modem to market. The mere fact that Kepler.M is so much more powerful/feature filled/energy efficient than anything else previously available in a Tegra product, and the mere fact that Kepler.M is coming to market earlier than expected is probably part of the reason why Tegra 4 was passed up by some vendors (with the other part of the reason being the strength of products such as S600 and S800).

 

[Alexko]

It is unfortunate that Tegra will be well off the pace this fiscal year vs. last fiscal year. Tegra revenue for the entire fiscal year will likely be close to ~ $500 million across all segments (auto, embedded, tablets, smartphones, etc).

That seems rather optimistic.

With about ~150M in the first half, it would require 350M in the second one.

 

[ams]

Not at all. Q1 FY14 and Q2 FY14 both reflect fiscal quarters where T3 production is steadily ramping down at the same time that T4 production is not ramping up (due to the delay in T4). Q3 FY14 and Q4 FY14 should both be much stronger quarters for Tegra relative to Q1 and Q2 for a few reasons: 1) T4 production steadily ramps up starting Q3 FY14; 2) T4 production maximized in Q3-Q4 FY14 as partners prepare to ship more products in volume during the holiday season; 3) T4i production steadily ramps up starting [end of] Q4 FY14.

Last but not least, NVIDIA has guided on full year Tegra revenues for Fiscal Year 2014 relative to Fiscal Year 2013. Fiscal Year 2013 had Tegra revenues of ~ $750 million across all lines of business. Fiscal Year 2014 will have Tegra revenues of ~ $500 million across all lines of business (with a margin of error of ~ 10%).

 

[Silent_Buddha]

I think that's highly optimistic of them. For FY2013, Tegra benefitted from the introduction of Windows RT as the only supported ARM chip. I doubt any OEM is going to be pushing it this year. And there's been no word from MS about whether they will push Windows RT this Holiday season and if they do, which ARM chip they'll be using.

Regards,
SB

 

[Alexko]

Plus, I think Tegra 3 is still selling. In fact I assume that most of the $50 million figure for Q2 was Tegra 3, since this quarter ended on July 28 and there wasn't much going on for Tegra 4 back then. But now the new Nexus 7 is out and I don't expect the old one to sell much, if at all. That means revenue generated by Tegra 3 should quickly drop to near zero.

To be honest I'm not even sure Tegra will generate more revenue in Q3 than Q2, but in any case I'd be really surprised if it reached $500 at the end of the year. I suppose a highly successful, Tegra 4-powered Microsoft tablet is still possible. Beyond that, all I see is a few design wins from OEMs that have never been successful in the tablet market. And maybe Asus, I think there's a new Transformer in the works, or possibly already out? I'm not sure.

Tegra 4i might be a different story, but a 2014 story.

 

[ams]

Try to think outside of the box. Tegra is much more than just tablets. Tegra Automotive revenue is expected to double over the course of this fiscal year. In this segment in particular, prior gen Tegra shipments do not simply drop off a cliff when next gen Tegra is ready. Considering that T4 production is ramping up starting Q3 FY14, and considering that Q3 historically is one of the strongest quarters of the year for Tegra due to the need to ship before the holiday season, it would be crazy to think that Q3 Tegra revenue will stay flat relative to Q2 Tegra revenue in FY14. Q2 is undoubtedly the trough this fiscal year for Tegra.

 

[Kaotik]

Tegra 3 is probably selling, but it's price was pushed to the gutters ages ago, and price is literally the only real factor it can compete with, if I haven't understood things wrong

 

[A1xLLcqAgt0qc2RyMz0y]

As for Nvidia, since the GPU maker is seeing weaker-than-expected profits, the company is unlikely to launch a price cut immediately, but will observe Hawaii's performance before making a move, the sources noted

Where exactly do they find these SOURCES as they do not seem to have a clue and just make up stuff.

for 2nd Quarter 2014
GPU revenue at $858.6 million up 9.3% Q-Q and 7.5% Y-Y.

Record GAAP and non-GAAP gross margins of 55.8 percent and 56.3 percent, respectively.

The GPU business continued to grow, driving our fourth consecutive quarter of record margins

http://nvidianews.nvidia.com/Releas...sults-for-Second-Quarter-Fiscal-2014-a05.aspx

http://phx.corporate-ir.net/External.File?item=UGFyZW50SUQ9MTk2NzE2fENoaWxkSUQ9LTF8VHlwZT0z&t=1