NVIDIA Tegra Architecture

But I think MS Sync was on Ford for several years.

Heard it wasn't very good but MS advertised it a lot. Just about every podcast was sponsored by Sync for about a year.
 
So, NVIDIA gets great automotive design wins that are shipping in volume *and* they don't have to waste a ridiculous amount of money on advertising? I feel really bad for them :(
 
Last edited:
Are they shipping lots of units?

The most high profile one in the past year seems to be the Nexus 9, which doesn't seem to be a huge seller.

Unless NV SOCs are in every Audi being sold these days, or there are some no-name Chinese devices that we're not that familiar with, where is this volume?

Last two quarters they're reporting incrementally higher revenues compared to the year-ago quarters:

http://nvidianews.nvidia.com/news/n...al-results-for-fourth-quarter-and-fiscal-2015

http://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-first-quarter-fiscal-2016

They tout the Automotive products under "Highlights" section of the press release for both quarters, not much specificity about mobile products, though they do talk up the desktop GPU products. For instance, these are the Highlights from the most recent quarter:

First Quarter Fiscal 2016 Highlights

During the first quarter, NVIDIA achieved progress in each of its platforms.

Gaming:

  • Launched NVIDIA® GeForce® GTX™ TITAN X GPU, the world's fastest GPU.
  • Introduced its first living-room entertainment device, NVIDIA SHIELD™, the world's first 4K smart TV platform that uses the Internet to connect you to a world of media, games, and apps.
Enterprise Graphics:

  • Introduced its roadmap for physically based rendering -- including Iray® 2015 rendering software; Quadro® M6000, the world's most powerful professional GPU; and the Quadro Visual Computing Appliance, with eight M6000 GPUs.
HPC & Cloud:

  • Drew more than 4,000 guests to the sixth annual GPU Technology Conference, which focused on deep learning and self-driving cars, and featured talks by leaders in these industries.
  • Major consumer web and cloud companies -- including Baidu, Facebook, Flickr/Yahoo, Microsoft and Twitter -- are using NVIDIA GPUs for deep learning.
  • Announced that its next-generation Pascal™ GPU architecture will accelerate deep learning applications 10X beyond the speed of current-generation Maxwell™ processors.
Automotive:

  • Announced availability of NVIDIA DRIVE™ PX, a powerful car computer that utilizes deep learning to enable self-driving capabilities.
- See more at: http://nvidianews.nvidia.com/news/n...irst-quarter-fiscal-2016#sthash.BAwZJrHT.dpuf

Maybe it's reading between the lines but if they had any hit mobile device products, they'd be touting them in these highlights.
 
Well, Audi is part of the VW group. A good quantity of Audis are sold but if you also factor in VW, Seat, Skoda, Porsche, Bentley, Bugatti and Lamborghini, you can see there is potential for a lot of revenue/design wins there in the longer term.
 
Well, Audi is part of the VW group. A good quantity of Audis are sold but if you also factor in VW, Seat, Skoda, Porsche, Bentley, Bugatti and Lamborghini, you can see there is potential for a lot of revenue/design wins there in the longer term.

Well, It is not because thoses brands are in the same group, that mean every brands will use the same hardware, supplier ( specially in the car market, where brands still got their own suppliers, their own engineers team and for many are just "acquisition " ) in addition, cheap actual models will not use any embedded advanced technology depending the range.

Then, again, it is not because they buy it today that brands will not move to a different supplier then. ( as for many supply on this markets, there's a lot of things who can enter in consideration: first will be the price / volumes, then how price and supply can be integrates in a same lineup, and cover every prices of a range of cars, not only one high end part ).
 
Yeah you wouldn't think VWs will have the same GPS and screen that a $60k Audi will have.

But they could have some lower end Tegras, because honestly, you don't need too much 3D graphics in a car system.
 
I know for a fact that Audi, VW, Skoda and Seat have shared quite a number of components between certain models in the past. This is from experience of driving in/travelling in different cars where many of the controls were pretty much the same! Not too surprising when the cars share chassis/engines and so forth.

As always with newer technology (ABS, for example), it tends to find itself in the higher end models first but then quickly works its way down the food chain, so to speak. No reason I can think of that this won't be the case for SOCs running the displays in cars. Even if NV aren't making much money there at the moment, if they've got their foot in the door at Audi, they could be set to expand their business enormously in this area.
 
Are they shipping lots of units?

The most high profile one in the past year seems to be the Nexus 9, which doesn't seem to be a huge seller.

Unless NV SOCs are in every Audi being sold these days, or there are some no-name Chinese devices that we're not that familiar with, where is this volume?

Last two quarters they're reporting incrementally higher revenues compared to the year-ago quarters:

http://nvidianews.nvidia.com/news/n...al-results-for-fourth-quarter-and-fiscal-2015

http://nvidianews.nvidia.com/news/nvidia-announces-financial-results-for-first-quarter-fiscal-2016

They tout the Automotive products under "Highlights" section of the press release for both quarters, not much specificity about mobile products, though they do talk up the desktop GPU products. For instance, these are the Highlights from the most recent quarter:

- See more at: http://nvidianews.nvidia.com/news/n...irst-quarter-fiscal-2016#sthash.BAwZJrHT.dpuf

Maybe it's reading between the lines but if they had any hit mobile device products, they'd be touting them in these highlights.

Again their automotive design wins are low volume/high margin. Since you are quoting financial results you can have a look at Tegra related revenues and think how many units those could represent in the end. If NVIDIA would had been hunting after high volumes they should had turned their strategy eons ago into something Mediatek alike. I might personally still believe they should have, but if they get with just one automotive package sold as much as Mediatek with a few handful of SoCs then they're definitely not at a disadvantage in the automotive per se.

Overall yes they don't make much if any profit from Tegras, but then again I don't recall them making any profit from past IGPs either where it was more to have a market presence than anything else.
 
I don't think it's wise to overestimate the per-chip revenue of automotive SoCs. It's not as high as anyone would love to believe.

I think the assumption was that the chips themselves might not be very expensive, but that lucrative platform/software contracts might accompany them.
 
Just curious how long they could stay in the game. TI and others bailed, even though growth in mobile devices have been very strong.

But apparently they're shut out of phones. Or they can't produce product at prices to compete with the likes of Mediatek on the one end and Qualcomm on the higher end.

Phones have the sheer volumes so limiting themselves to tablets and cars is an interesting strategy. If car electronics became lucrative enough, you can be sure Qualcomm and others which they couldn't compete against in phones will enter the car market.
 
I don't think it's wise to overestimate the per-chip revenue of automotive SoCs. It's not as high as anyone would love to believe.

I wasn't thinking of just chip revenue, but revenue in total. To avoid further brainfarts: is NV just selling SoCs to car manufacturers or are their other costs for support and/or sw also? If yes for the latter I had the entire enchilada in mind hence me mentioning "package".
 
Depends. In the automotive space it's rare for a silicon vendor to deal with the car vendor directly. There's usually a module supplier/ODM in the middle. So it all depends if they're also gearing up to be their own module supplier as well. That's a massive cost to take on, because the module supplier takes on the burden of certification and fit-for-purpose proofs, longevity testing, pre-physical integration, etc.

Support and software is therefore structured differently because of the middle man aspect. It's not going to be large amounts of revenue for NV to book if they're not the module supplier too, but being the module supplier isn't anything you can do in a short period of time.
 
Well, Audi is part of the VW group. A good quantity of Audis are sold but if you also factor in VW, Seat, Skoda, Porsche, Bentley, Bugatti and Lamborghini, you can see there is potential for a lot of revenue/design wins there in the longer term.

That a good point! In addition, at the 2015 Detroit Auto Show they also showed Tegra-Powered VW, Audi, BMW, & MINI's. They mention the newest addition is the Golf whose new infotainment system is Tegra powered. Audi has a new option available "Tegra-powered" tablet for passengers which is hard-wired into "Audi’s next-generation in-car infotainment systems. So passengers can use it as a remote for the car’s navigation, audio, phone and other functions." They also mention BMW 6 series is also powered by the Tegra as well as the Mini Countryman.

And Tesla has been using the Tegra since 2012..

http://blogs.nvidia.com/blog/2015/01/15/motown-tegra/

Multiple Tegra modules in each car ...
Tegra also helps drivers and passengers take command of Audi’s new flagship SUV, the Q7, which will sport as many as four NVIDIA Tegra processors.
http://blogs.nvidia.com/blog/2015/03/11/audi-geneva/

After accepting the 2015 PACE Award for the Tegra Visual Computing Module, Nvidia mentioned they will be shipping their car computers to carmakers ...

"Our NVIDIA DRIVE car computers use the newest version of the VCM. They’ll start shipping to carmakers and developers next month. NVIDIA DRIVE computers are the world’s most advanced systems for self-driving cars and digital cockpits."
http://blogs.nvidia.com/blog/2015/04/21/pace-award-tegra-vcm/

"In Nvidia’s (NVDA) Tegra business, the automotive segment is the fastest growing subsegment. Apart from being a fast growth segment, it offers higher gross margins as compared to the devices. In 4Q15, Nvidia’s automotive segment revenue almost doubled on a year-over-year basis."

"According to Nvidia, the company’s automotive platforms appear to be on a sharp upward trajectory. In 2015, over 7.5 million cars are using Nvidia’s technology, up from 4.7 million in 2014. For the past decade, Nvidia has been working on building its automotive computing platform, and therefore is in a strong position to leverage this growth. The company’s customers include prestigious automobile companies like Volkswagen, Honda (HMC), BMW (BMW), and Audi (AUDVF)."
http://marketrealist.com/2015/02/potential-key-drivers-nvidias-tegra/
 
NVidia seem to be a bit sneaky here. They have attached a Mobileye’s chip to many of there own car solutions and its really the Mobileye’s doing the driving not the Tegra chip for many versions. The Audi prototype that recently completed a 550 mile piloted drive from Silicon Valley to Las Vegas was Mobileye.
its a Tegra K1 SoC with a Mobileye’s EyeQ3 processor.

Audi, BMW, Fiat, Ford, General Motors, Honda, Nissan, Peugot, Citroen, Renault, Volvo and Tesla all seem to be going with Mobileye not leaving a lot of room for Tegra. Can NVidia compete against a company's dedicated to automotive solutions? Will Audi go with EyeQ4 or TegraX1? Perhaps a mix?
 
Audi, BMW, Fiat, Ford, General Motors, Honda, Nissan, Peugot, Citroen, Renault, Volvo and Tesla all seem to be going with Mobileye not leaving a lot of room for Tegra. Can NVidia compete against a company's dedicated to automotive solutions? Will Audi go with EyeQ4 or TegraX1? Perhaps a mix?

Do you have url's that supports your views?

Edit: While Nvidia is a newcomer to the automotive market, sales and trends from the url's above suggest automotive manufacturers are more than willing to try different sources for their automotive computer solutions.
 
Last edited by a moderator:
Do you have url's that supports your views?

Edit: While Nvidia is a newcomer to the automotive market, sales and trends from the url's above suggest automotive manufacturers are more than willing to try different sources for their automotive computer solutions.
Are NVidia really a different source when they are using the same source as everyone else just trying to pass it off as their own? As far as I can see all those Tegra 3, 4 and K1 automotive solutions are not really NVidia technology nor NVidia software. It’s all Mobileye’s EyeQ3 hardware with Mobileye’s software run via MIPS CPU’s not Tegra. It looks to me like NVidia are just using marketing PR to pass off someone’s else automotive solutions as their own. Please correct me if I am wrong. How can they say Tegra automotive when the automotive stuff is not done on the Tegra chip or Tegra software?
 
Why is Tegra more likely to be adapted than Snapdragon by car makers?

Is it hardened for temperature and moisture extremes or something?

You also wonder how many people will bite on built-in tablet in cars if they already have phones and tablets. Why overpay for car electronics when in a year or two, you can get a much better device for $200-300.

And every year, the devices will be faster while the electronics you bought for the car is stuck.
 
Why is Tegra more likely to be adapted than Snapdragon by car makers?

Is it hardened for temperature and moisture extremes or something?

You also wonder how many people will bite on built-in tablet in cars if they already have phones and tablets. Why overpay for car electronics when in a year or two, you can get a much better device for $200-300.

And every year, the devices will be faster while the electronics you bought for the car is stuck.

I would imagine it has a lot to do with nVidia's work in convolutional neural networks. I doubt snapdragon can get close, and the software certainly isn't there. If nVidia is giving a pretrained network that can detect objects accurately to car makers, then that is a competitive advantage for sure.

It is very likely that convolutional neural networks will be a driver of compute capabilities in the future, similar to how graphics were also a driver. How long until nVidia starts delivering chips built specifically to run neural networks? (It probably won't be anytime soon since the techniques used are changing so rapidly.)
 
Back
Top