http://blogs.barrons.com/techtraderdaily/2016/03/18/mobileye…
This announcement from Mobileye I thought was pretty insightful.
Firstly it touches on the path to full autonomous driving which incorporates:
25 manufacturers on board
Visual, Radar and Lidar technology
Proprietary roadmap and REM intelligence building
Cognitive computing and neural intelligence
Secondly it highlights some of the criticalities that include:
Density of processor cores
Super computer processing
Performance to power envelope
Highly advanced node fabrication
Data bandwidth and transfer speed
Thirdly it gives us an indicator of potential winners:
Mobileye
STMicro
ARM who are in both of the above
Potentially Skyworks
and suggests potential losers in this market as a result:
Intel
Ambarella
Qualcomm (except potentially for modems)
Qorvo (who don’t necessarily work with STMicro)
Cheers
Ant
5 Likes
Thirdly it gives us an indicator of potential winners:
Another potential is NXPI.
And yet another is NVIDIA, a strong player in autonomous vehicles, data centers, and of course video gaming. Multiple possible futures and very strong earnings a few days again. I think its 1YPEG was too high for this board so I have resisted talking about it too much.
And yet another is NVIDIA, a strong player in autonomous vehicles, data centers, and of course video gaming. Multiple possible futures and very strong earnings a few days again. I think its 1YPEG was too high for this board so I have resisted talking about it too much.
I, for one, would like to hear more about it, if you are so inclined.
2 Likes
Yeh there are plenty of additional players that could be considered for this market, I was restricting my comment on winners and losers to those that were implicated in Mobileye’s press release. Given that they have a massive first mover advantage, 25+ car makers on board and a huge moat (with their integrated platform, mileage database and client crowd-sourced road mapping); I would say those that are in the Mobileye’s allied platform stand a greater chance of success.
There are plenty of other players including NXP, Nvidia, Bosche, Cypress, CAMP, Sierra, Telit etc connected to smart cars and AD but the ones identifiable in mobileye’s plans and their excluded competitors are the core ones I mentioned.
In particular I think it strengthens STMicro vs Qualcomm, MBLY vs AMBA and ARM vs Intel and vicariously Skyworks vs Qorvo & Avago (I need to confirm that last one a little more - but it is definitely better for Skyworks if Qualcomm doesn’t win out as per the Samsung S7 smartphone experience).
It is looking increasingly like another mobile processor battle outcome in smart cars for ARM vs Intel. I knew bandwidth and data speed was important but I hadn’t heard such a telling requirement for a performance vs power envelope. If that is the case and power is critical - that is going to help ARM vs Intel’s Quark & Atom in cars just as it helped ARM vs Intel’s Atom in mobile.
Cheers
Ant
1 Like
http://blogs.barrons.com/techtraderdaily/2016/06/10/mobileye…
Another collateral article from the same publication.
MBLY is said to have a “quasi” monopolistic position in visual related autonomous driving and a large and ever growing user base.
That is the catch though, what if visual is succumbed by disruptive radar or such…yes, well that is always the rub, all the time. AOL fell when dial up caved to dal.
However, absent unibiquitous road sensors, seems difficult to me to envision how a car is suppose to drive itself based upon radar, or satellite without visual.
but what you gonna do.
Tinker
1 Like
That is the catch though, what if visual is succumbed by disruptive radar or such…yes, well that is always the rub, all the time. AOL fell when dial up caved to dal.
I wonder. Visual has two components, image capture and image analysis. I would think that MBLY advantage rests in the latter which could be adapted to other kind of imagines. On the other hand their visual database could become obsolete.
For me technology is mostly in the too-hard file. 
Denny Schlesinger
A couple of points:
-
There are “only” 60 million cars made each year. Contrast that to over 1 billion smartphones and tablets in 2014 , as reported by Gartner (http://www.gartner.com/newsroom/id/2996817). Worse, the vast majority of those vehicles don’t have any ADAS equipment. Even when the US mandates some features (2021 I think), that’s only 14 million vehicles at most (that includes trucks, for instance). So, there’s little money in chips for cars, at least compared to phones.
-
Systems and sub-systems are another story. Hardware for things like infotainment can easily reach $1K for the high-end, although somewhat dependent on volumes. Companies like Visteon, Delphi, Bosch, Continental, etc. do well building custom variations on base technology for different OEMs.
-
And then there’s the software. In my view, that’s where the value and profits lie. And that’s mostly at the Tier 1s mentioned above.
-
MobileEye probably doesn’t have road data from automobiles. First, most automobiles don’t capture continuous location data and beam it back to HQ. And even if they did, would the OEM share that with Mobileye? Tesla is reportedly doing just that, but there’s no indication of how much is shared.
-
Right now Mobileye is focused on visual/camera based ADAS. That has benefits in terms of required equipment, but also has related limitations. Can the camera see through rain better than a driver? Can the system know where the lanes are if the lines are covered with snow?
I haven’t invested in any ADAS plays yet because I can’t tell the winners from the losers, even at the macro business model level, much less an individual company level.
4 Likes
The most likely event is that more than one type of sensor will be used. Though I expect cameras to predominate. Thanks to mobile phone development they are cheap.
Humans control their cars mostly by vision so we know that works. Most of the time.
The real key is interpretation of the data, software.
And perhaps how that software is integrated into the mechanics of the vehicle . I am unsure whether this will be designed in by the car companies or bought from another company, a supplier.
The history of infotainment systems in cars suggest that car companies are not likely to do this too well on their own.
At this stage it is hard to even design a basket of stocks.
1 Like
Hi Smorgasbord1
Just a few thoughts re your points:-
-
Mobileye’s functionality can be and is retrofitted into existing cars as well as new cars
-
Infotainment systems aren’t mandatory regulated systems like driving assistance.
-
Mobileye’s killer advantage is in the software - the chips and hardware could come from anywhere
-
Mobileye has millions of miles of road data - the most of any player and it is from automobiles. It also has a shared gathering model to build more
-
Mobileye has been testing radar and lidar as well as visuals for years and its next system will build in that incorporation. They have said the solve needs to be a mixed model. You can’t have visual from one company, lidar from another and radar from another so with their first mover advantage they should have the the opportunity to consolidate on their platform.
A
3 Likes
Mobileye has millions of miles of road data - the most of any player and it is from automobiles. It also has a shared gathering model to build more
Can you tell me more about this? Do Mobileye systems have built-in cellular connectivity, or do they rely on the car’s own connectivity? Are OEMs collecting the data and sharing it with Mobileye, or is Mobileye collecting it themselves with permission of the OEM?
I’ve read that Tesla is doing this and literally play “what if” scenarios inside their systems to see if what the car would do matches what the driver did, or what should have been done. And, we know that Tesla has the ability to collect real time GPS and car status data - but is Tesla, for instance, sharing that data with Mobileye or keeping for their own program that happens to include some Mobileye technology?
Yep Tesla is a game changer and I don’t think people realize just how much and I am not even talking about the electric motor.
How about 18 million miles of data -
"WSJ: Mobileye makes extensive use of “tagged” videos of test rides—manually annotated videos showing where cars are during a driving sequence, as well as things like road signs and obstacles. How does the company use this data?
Mr. Shashua: We have access to a lot of data. Tagging data is part of the preparations needed in order to enable machine learning. This is how the algorithms learn. We have tagged videos of rides for 30 million kilometers [about 18.6 million miles]. This data comes from lots of geographies—Europe, the U.S., Asia. It’s also diversified in terms of visibility conditions and road conditions—it’s really unbiased, and that is something that is hard to produce. We have 500 people in Sri Lanka [annotating videos manually] and a few dozen in Israel. It takes time to create data. The fact that we work with so many auto makers helps. Just one company would have found it hard to produce that amount of data."
http://www.wsj.com/articles/q-a-mobileye-founder-talks-self-…
and
“Mobileye is a leader in advanced driver assistance systems (ADAS), whose chips and software take in data from car cameras. Mobileye has introduced a software system called Road Experience Management, which collects data for mapping purposes. REM collects major landmark and road information every kilometer…“GM is also working with Mobileye on their REM localization product. Once fully deployed, this will allow GM to collect 100 million miles of data per day in the U.S. alone,” Spak wrote… Aside from GM, Mobileye has signed up Volkswagen (VLKAY) and Nissan (NSANY) for its REM program.”
http://www.investors.com/news/technology/mobileye-to-race-te…
That’s just the first two links when I googled…
https://www.google.com.sg/?gfe_rd=cr&ei=B7xeV47WO86AoAPc…
then there’s this…
http://electrek.co/2016/01/08/mobileye-tesla-mapping/
Ant
4 Likes
One thing that amazes me is the Nav computer. It has practically every road and exit mapped, and graphically looks rather accurate on the dashboard screen.
Autonomous driving is of course more complex, requiring even more detailed information, but it is not wholly different, just a matter of more data needed, and more dynamic.
Multiple firms provide Nav data. I am sure that multiple firms will be able to provide autonomous driving data as well.
When Mapquest first came out, it was like magic, it was just amazing. Now we are of course way past that.
Tinker