NVidia - how is it going as a company?

[KingFatty]

I was surprised this past weekend to see an NVidia rep inside a Microcenter store. He was trying to sell the shield gaming thing that has Tegra 4 in it.

But in the store, there were teens and young adults crowded around another gaming system, I think it was just a plain old PS3 or Xbox 360 fighting game. Nobody seemed to be interested in the Shield.

So I thought hmmm, is this desperation by Nvidia to try to hire their own people to try to sell the Shield, but is the shield doomed? The rep claimed that sales were great and they were selling out etc., but I was the only one paying any attention to him. I just wonder maybe I got the wrong vibe in that particular Microcenter.

Then I'm thinking is NVidia OK? Did they bet big on Tegra 4 and Titan, but is the future OK for them? For Shield, maybe it's just an experiment and a way to try to sell more Tegra 4 chips. But for Titan, I wonder how much longer PC sales will be a good revenue source? Will companies like Samsung and Apple get all the sales that previously went to PC manufacturers and dedicated hand-held consoles? I dunno, I'm just starting to worry about companies that are pushing sales for things that seem like it is no longer the future? How can Nintendo even stay relevant? I just hope Nvidia is not trying to copy Nintendo.

But looking big-picture, can someone talk about how Nvidia has a very good business, perhaps with non-gaming PC equipment that will be needed for rendering movies etc., or non-handheld gaming equipment, just a good future for them?

 

[Granseth]

The new Asus Transformer 701 is using Tegra 4, but I have no idea if it will be selling well.

But if I where to buy the Shield device, would I be locking myself to nVidia hw in my desktop or would the streaming from PC work with AMD GPU as well?

 

[NTMBK]

They probably "sold out" in the same sense JHH talks about here: http://techreport.com/news/25210/nvidia-early-shield-sales-have-been-great

Everything that we shipped so far has sold out and – but we're just starting to ramp production, we've shipped out to our partners only several thousand units

 

[NTMBK]

But if I where to buy the Shield device, would I be locking myself to nVidia hw in my desktop or would the streaming from PC work with AMD GPU as well?

Locking yourself to NVidia. The Shield streaming uses the hardware encoder block built into Kepler GPUs to keep latency down.

 

[f1sherman]

Shield is a $10M bet, i.e. 1% of R&D.

JHH son could have financed that with his lunch money.

 

[Ibra]

And how's AMD doing? Sending minions all over the forums trying to get APU design wins? :hmm:

Infraction issued for thread crapping. Enjoy your vacation.
-- stahlhart

 

[raghu78]

k. lets split it into 4 segments.

1. Consumer GPU market - desktop and notebook. Here nvidia is doing well on the strength of its Kepler products. The big concern is a shrinking dGPU market especially at the low end which has high volumes. Intel with Haswell / Broadwell and AMD with Kaveri are going to accelerate the eventual demise of the entry level GPU market(below USD 100). Both these companies are selling powerful iGPUs. Once stacked DRAM using TSV (thin silicon vias) becomes available the bandwidth problem too should be solved. So it means less discrete GPU revenue and less money for funding future GPU R&D.

2. HPC and professional . Here Nvidia dominates on the strength of their Tesla and Quadro products. Since they are at 90% marketshare, Nvidia has everything to lose. Intel and AMD are both making efforts to take a portion of this market. Intel with Knights Landing and their x86 based HPC products. AMD with Hawaii and their Firepro products and with Berlin APU based HPC clusters. Nvidia's strength has been CUDA and their software driving these products. Longterm Nvidia will have to cede marketshare. They will be a dominant player but not a monopoly with 90% share. Nvidia's Denver 64 bit ARM architecture is key to their long term leadership in this market as eventually HPC clusters built out of single chip SOCs with Terabytes of bandwidth using stacked DRAM will dominate. Nvidia needs a strong CPU architecture to compete with Intel / AMD.

3. Mobile - Smartphones and Tablets. the problem here is the phone market is overcrowded. Many former giants like Texas Instruments have quit this market. Nvidia is feeling the same problem. their Tegra revenues have crashed in H1 2013. Also the top 3 - Samsung, Apple and Qualcomm eat up roughly 90% of the market and are slowly killing the smaller companies. Intel is getting into this market in a big way with Baytrail. So it gets worse from here on. its only a matter of time before only 3 -4 players are left. The rest will wither away. I predict Qualcomm, Intel, Apple and Samsung will be the last companies left standing. As for Shield there is not much hope for it.

4. GPU licensing and gaming market- for the next gen of PS4/XB1/Wii U Nvidia is out. so for the next 6 - 7 years their gaming revenue is just going to be negligible. PS3 related licensing revenue will keep trickling in but its going to be very small.

Nvidia has started to license its GPU IP to ARM SOC vendors. This could be considered a sign of acceptance by Nvidia that they don't feel confident of growth in a crowded mobile market selling ARM SOCs but see huge potential through GPU licensing. This could be a game changer. Nvidia could get Apple or Samsung to license their GPU designs for use in mobile SOC. currently Imagination and ARM license their GPU IP - PowerVR and MALI. Qualcomm designs their own GPU IP. Apple licenses Imagination's GPU designs and Samsung licenses ARM's GPU designs.

So overall Nvidia has a lot of challenges and few opportunities. They need to reinvent themselves to remain as a leading technology company over the next decade and further.

 

[Grooveriding]

Shield is a joke, probably a way to get rid of all those Tegra 4 chips no one is buying , but that doesn't mean there is anything wrong with the GPUs they are selling at massive markups.

Shield is like a $150 item at best, the idea that they thought it warrants a $300 pricetag is laughable. I saw that article about them 'selling out' that then went on to clarify they only released 5000 units to each of the official vendors at launch.... lol. Who on this forum has one ? That didn't get one for free ?

 

[NTMBK]

They will be a dominant player but not a monopoly with 90% share. Nvidia's Denver 64 bit ARM architecture is key to their long term leadership in this market as eventually HPC clusters built out of single chip SOCs with Terabytes of bandwidth using stacked DRAM will dominate. Nvidia needs a strong CPU architecture to compete with Intel / AMD.

Don't forget that NVidia and IBM announced a partnership recently to more tightly couple POWER CPUs and Nvidia GPUs. What that actually results in remains to be seen, but it certainly seems like a credible plan B for HPC in case Denver doesn't pan out.

 

[seitur]

Currently it is doing alright. Closest (few years ahead) future is relatively secure as well. More distant future is muddy as market for discrete graphics will slowly decrease.

 

[raghu78]

Don't forget that NVidia and IBM announced a partnership recently to more tightly couple POWER CPUs and Nvidia GPUs. What that actually results in remains to be seen, but it certainly seems like a credible plan B for HPC in case Denver doesn't pan out.

http://www-03.ibm.com/press/us/en/pressrelease/41684.wss

thats interesting info. IBM's POWER is an excellent architecture for big iron servers. the fact that IBM has gone with an open development alliance based on IBM's POWER microprocessor architecture with Google and Nvidia as founding members bodes well for the entire tech industry. So we are going to have a strong alternative to x86 and ARM with OpenPOWER. :thumbsup:

 

[KingFatty]

k. lets split it into 4 segments.

1. Consumer GPU market - desktop and notebook. Here nvidia is doing well on the strength of its Kepler products. The big concern is a shrinking dGPU market especially at the low end which has high volumes. Intel with Haswell / Broadwell and AMD with Kaveri are going to accelerate the eventual demise of the entry level GPU market(below USD 100). Both these companies are selling powerful iGPUs. Once stacked DRAM using TSV (thin silicon vias) becomes available the bandwidth problem too should be solved. So it means less discrete GPU revenue and less money for funding future GPU R&D.

2. HPC and professional . Here Nvidia dominates on the strength of their Tesla and Quadro products. Since they are at 90% marketshare, Nvidia has everything to lose. Intel and AMD are both making efforts to take a portion of this market. Intel with Knights Landing and their x86 based HPC products. AMD with Hawaii and their Firepro products and with Berlin APU based HPC clusters. Nvidia's strength has been CUDA and their software driving these products. Longterm Nvidia will have to cede marketshare. They will be a dominant player but not a monopoly with 90% share. Nvidia's Denver 64 bit ARM architecture is key to their long term leadership in this market as eventually HPC clusters built out of single chip SOCs with Terabytes of bandwidth using stacked DRAM will dominate. Nvidia needs a strong CPU architecture to compete with Intel / AMD.

3. Mobile - Smartphones and Tablets. the problem here is the phone market is overcrowded. Many former giants like Texas Instruments have quit this market. Nvidia is feeling the same problem. their Tegra revenues have crashed in H1 2013. Also the top 3 - Samsung, Apple and Qualcomm eat up roughly 90% of the market and are slowly killing the smaller companies. Intel is getting into this market in a big way with Baytrail. So it gets worse from here on. its only a matter of time before only 3 -4 players are left. The rest will wither away. I predict Qualcomm, Intel, Apple and Samsung will be the last companies left standing. As for Shield there is not much hope for it.

4. GPU licensing and gaming market- for the next gen of PS4/XB1/Wii U Nvidia is out. so for the next 6 - 7 years their gaming revenue is just going to be negligible. PS3 related licensing revenue will keep trickling in but its going to be very small.

Nvidia has started to license its GPU IP to ARM SOC vendors. This could be considered a sign of acceptance by Nvidia that they don't feel confident of growth in a crowded mobile market selling ARM SOCs but see huge potential through GPU licensing. This could be a game changer. Nvidia could get Apple or Samsung to license their GPU designs for use in mobile SOC. currently Imagination and ARM license their GPU IP - PowerVR and MALI. Qualcomm designs their own GPU IP. Apple licenses Imagination's GPU designs and Samsung licenses ARM's GPU designs.

So overall Nvidia has a lot of challenges and few opportunities. They need to reinvent themselves to remain as a leading technology company over the next decade and further.

Thanks for the insightful post! I was thinking your #2 is a good position for NVidia, and seems like the consensus is that they'll be in good shape going forward, even if the consumer PC industry as a whole declines.

Just saw depressing article about the death of Nintendo on Techcrunch, but good to hear that Nvidia's shield can succeed or fail and not really affect Nvidia:
The Death Of Nintendo Has Been Greatly Under-Exaggerated
http://techcrunch.com/2013/09/04/nintendont/

I wish that something changes to where consumer PCs become more popular and we get a healthy discrete video card market again. I mean, why why why can't laptops have expansion slots for video cards? Or even tablets? It would be so awesome to have upgradable graphics on laptops that are standardized.

 

[stahlhart]

The member callouts and thread derailing are starting to get out of control -- not just in this thread, but in this forum in general. Either you guys stop it, or I will. Stay on topic.
-- stahlhart

 

[AnandThenMan]

Nvidia knows how to make money, not worried about them at all. I do think they should stop screwing around with Tegra and get serious, it's going backwards.

 

[VirtualLarry]

Nvidia knows how to make money, not worried about them at all. I do think they should stop screwing around with Tegra and get serious, it's going backwards.

Nvidia-branded autonomous vehicles, packed with quadro cards to do all of the processing.

 

[blackened23]

Nvidia knows how to make money, not worried about them at all. I do think they should stop screwing around with Tegra and get serious, it's going backwards.

Getting involved with ARM SOCs was the right move by nvidia - they have to look beyond the PC for potential income, because as we all know the PC market is declining.

That said, Tegra 4 was a massive mistake. They've lost practically every *good* design win that Tegra 3 had, and they're resorting to putting these chips in low volume SKUs that won't sell well. None of the current design wins will sell like the prior generation nexus 7, that was huge for nvidia. They just don't have that with Tegra 4.

If there's a lesson to be learned for Nvidia, here, it's that you can't treat the mobile market as the PC market. You can't come late to the party and just expect folks to line up for your product, if you do - you will get slaughtered. That works in the PC market, but not in the ARM SOC race. And that is precisely what happened with the Qualcomm snapdragon which is getting every design win worth winning. That's not very good news at all for NV since they haven't earned a profit with the Tegra division yet, but i'm pretty confident they won't make the same mistake with Tegra 5. It's a learning process for them. As well, they're obviously making tons of money from the PC and HPC markets. Nvidia isn't hurting despite the Tegra 4 setbacks. But, Tegra 5 will be a make it or break it type of deal for nvidia - if Tegra 5 fails to gain traction as Tegra 4, nvidia will have to think long and hard about their future in the mobile ARM SOC race. They need the design wins that move 30+ milliion units like the Nexus 7 - throwing SKUs out there like the shield that will move maybe 500k units just won't cover their R+D expenses in the long haul.

 

[wand3r3r]

I agree with some of your points. Another twist is that while apparently (some) PC fanboys will wait for their brand "designer" GPUs, nobody cares what processor is in their tablet, phone, etc. Being late to the party = lose the sales.

 

[ams23]

The investments that NVIDIA has made over the last few years in visual computing/mobile computing/baseband technology will start to bear serious fruit in the very near future. Mark my words, NVIDIA will be a significant force in the mobile computing space starting with Logan and moving forward with Parker and future SoC's (note that Logan is NVIDIA's first SoC using their most modern-day GPU technology, and is also their first SoC being designed in conjunction with or concurrently with their next gen mobile processor) . The key challenges for NVIDIA are to bring their baseband modem technology to market, and to bring high performance ARM processors to Windows.

Tegra 4 was late to market, but it still has some quality design wins (including various tablet or tablet-convertible products by HP, Asus, Toshiba, Bungbungame, Kobo, etc. as well as the new amazing Xiaomi Mi3 smartphone that will be sold through China Mobile, the world's largest wireless carrier). In the meantime, Tegra Automotive is doing quite well, with approximately 100% revenue growth this year (and continued 100% year-over-year revenue growth over the next two or three years). As for Shield, I actually purchased one, and it is ridiculously good for the niche that it serves (and I am not the only one who feels this way; read the Newegg customer reviews on this device, they are amazingly good). Shield is a device aimed at showcasing current and future generations of Tegra processors, and is one of many ways to push forward Android gaming. Shield 2 due out in 1H 2014 will have a fully modern GPU featureset and will have more GPU GFLOPS throughput than PS3 and Xbox 360 (and more than 8800 GTX too!), all in the palm of your hands. In the future, Shield will also be an NVIDIA GRID client.

 

[artvscommerce]

I think between what they're doing with GPU Computing / CUDA and their new GRID products, they will be just fine. I suspect in a few years they will be bigger than ever before.

 

[Imouto]

Expect the very same story with the Tegra 5 as happened with the Tegra 4. Qualcomm and Apple have the initial 20 nm production and Nvidia will launch it at 28 nm.

As a whole it will be healthy, it's just its Tegra division what's doomed.

 

[FearfulSPARTAN]

The investments that NVIDIA has made over the last few years in visual computing/mobile computing/baseband technology will start to bear serious fruit in the very near future. Mark my words, NVIDIA will be a significant force in the mobile computing space starting with Logan and moving forward with Parker and future SoC's (note that Logan is NVIDIA's first SoC using their most modern-day GPU technology, and is also their first SoC being designed in conjunction with or concurrently with their next gen mobile processor) . The key challenges for NVIDIA are to bring their baseband modem technology to market, and to bring high performance ARM processors to Windows.

Tegra 4 was late to market, but it still has some quality design wins (including various tablet or tablet-convertible products by HP, Asus, Toshiba, Bungbungame, Kobo, etc. as well as the new amazing Xiaomi Mi3 smartphone that will be sold through China Mobile, the world's largest wireless carrier). In the meantime, Tegra Automotive is doing quite well, with approximately 100% revenue growth this year (and continued 100% year-over-year revenue growth over the next two or three years). As for Shield, I actually purchased one, and it is ridiculously good for the niche that it serves (and I am not the only one who feels this way; read the Newegg customer reviews on this device, they are amazingly good). Shield is a device aimed at showcasing current and future generations of Tegra processors, and is one of many ways to push forward Android gaming. Shield 2 due out in 1H 2014 will have a fully modern GPU featureset and will have more GPU GFLOPS throughput than PS3 and Xbox 360 (and more than 8800 GTX too!), all in the palm of your hands. In the future, Shield will also be an NVIDIA GRID client.

Sorry but iv heard stuff like this before with tegra but I just dont see it happening, nvidia losing the nexus 7 was the killer for tegra 4 imo. Plus what other meaningful design wins do they have? They already said they will see a moderate to large drop in tegra revenue year over year, I just dont see logan saving them next year, people said that about tegra 4 this year...

 

[AnandThenMan]

Tegra is always one step behind. And when you're behind, you essentially have to intro two generations in the same cycle as the market leader pulls one. Very difficult to do.

 

[ams23]

@ FearfulSPARTAN: The Xiaomi Mi3 Tegra 4 design win with China Mobile is pretty big news for Tegra 4 for a variety of different reasons. Xiaomi is one of the fastest growing companies in history ($10 billion valuation after only three years in business). China Mobile is the world's largest wireless service provider. And the system performance of the 1.8GHz Tegra 4 variant in this smartphone form factor appears to be very good even compared to the 2.3GHz S800-8974AB variant (even though the latter does have better GPU-centric benchmark performance):

Antutu: 36582 vs. 36026
Quadrant: 18702 vs. 18975
Geekbench2: 4106 vs. 4139

So what is the difference between Tegra 2/3/4 vs. Tegra 5 "Logan" and Tegra 6 "Parker"? The difference is that, with Tegra 2/3/4, there was only one primary design team working with a less-than-modern GPU architecture. So, after Tegra 2 was done, they started Tegra 3, and after Tegra 3 was done, they started Tegra 4 (all the while having strict SoC die size area requirements to keep production cost low, and all the while being unable to heavily leverage existing modern-day GPU assets). Starting with Tegra 5 "Logan" and Tegra 6 "Parker", NVIDIA is designing two modern-day SoC architectures at the same time. This will make a big difference in both performance (with a modern-day GPU feature-set) and time-to-market.

When Tegra 5-powered devices first come to market (which should be as early as Mar-Apr 2014 since the "Logan" SoC started sampling in ~ June 2013), the peak GPU performance of Kepler.M will be unmatched in the ultra mobile space in my opinion. I do not believe that this level of peak GPU performance will be matched in this space until a somewhat mature 20nm fabrication process in the second half of 2014. And right around the corner in 1H 2015 will be Maxwell.M GPU and the circle will turn again in NVIDIA's favor. So instead of being one step behind with respect to GPU performance in the ultra mobile space, they will be one step ahead starting with Tegra 5 "Logan". The writing is on the wall so-to-speak.

 

[Granseth]

I will be very impressed if nVidia manage to have a shader that is efficient in everything from a mobile soc to a high performance gpu in a supercomputer.

 

[ams23]

If my forecasting skills and intuition are correct, I believe that Kepler.M will revolutionize the ultra mobile GPU space. We will see in a few months!