I think the fear is (and I admit I know very little about MBLY the company) that if Tesla shows that radar based assisted navigation is superior to the camera based solution that MobilEye currently offers and MobilEye has nothing else to offer that someone else might come along and scoop up (some of) MBLYs market share with a radar based solution.
Maybe MBLY is quick on their feet and can adapt quickly with no negative effects. After all, this is all software and programming since what Tesla is doing is utilizing the same hardware.
It is not that radar is superior, I think. It is that radar interpretation is grotesquely complex, difficult and place-sensitive. Absent something approaching "thick data" on every road, radar will be quite error prone. Thus, cameras and other detection technologies will continue to have a place. Cameras are better in optimal visual conditions (surprise
) but are woefully inadequate in adverse weather and difficult terrain. There is no surprise that the Singapore and Pittsburgh autonomous efforts mimic the Google Palo Alto efforts. In each such case there is very Thick Data and the autonomous operation is limited to only those routes where Thick Data is available.
Tesla offers something in between, but crowd-sourced data collection is allowing some tangible benefits approaching the Thick Data models. How Tesla is actually managing that is something I, for one, think is really amazing but I do not pretend to understand how they're doing it.
With all the global efforts, MobilEye is just one player, albeit an early one that has had considerable success. They have emphasized processing and cameras, plus the early mover advantages. We now have legions of cartographers, satellite mapping data mavens and big data specialists> There are also a few geologists who specialized in seismology, a notoriously difficult arena that hunts for tiny valuable signals amidst mostly loud and irrelevant noise. This will all play out to our benefit IMHO, with Tesla continuing to lead in real-world application of these techniques.
As the Tesla advances become more evident in this arena some of the drag on price application will dissipate. No doubt GAAP positive earnings will too. Sometime people will begin to perceive that Tesla is profitable on a steady state basis now, but teh high growth rate will penalize GAAP for a long time to come. Further Tesla is carrying inventory until actual end-cutomer sale, which no other car manufacturer does and costs capital, as does the proprietary distribution system and the charging network. All that that is paying off, but it does eat tons of capital when rapidly growing. Just imagine the steady state!
Vertical integration has many benefits and a few disadvantages too. This biggest disadvantage is supporting high growth. Somehow the financial press and analysts cannot see that point and make some adjustments to consider apples to apples. BMW might well have 8% ROS go slightly negative were they vertically integrated as Tesla is.
