Yesterday’s big TSLA-related news IMO is the announcement that future FSD versions will be vision only. Reasons why this is a big deal:
1. Radar already seemed redundant. Humans don’t use radar to drive.
2. Future cars will probably see radar systems removed. Benefits of this:
a. Increased margin per car on the order of a few hundred dollars, or cheaper cars. Either way, a win.
b. Less parts = more reliability.
c. Less power consumption, both for the radar unit itself, and for the processing of this data. Also, more compute time available to process vision data.
3. Passive vision doesn’t get confused by bounced reflections (like when driving thru a tunnel).
4. Shows that Tesla has high confidence in the reliability of its vision sensing.
5. AP team can focus on vision 100%. In a way, narrows the scope of the problem.
The only possible downside of eliminating radar that I can see is that you can’t bounce passive vision under the car in front of you to see the car in front of that one. You may be able to get around this with vision by training the system to identify cars by looking through the lead car’s windows. Tesla might find they want to keep radar for this reason, but I’m betting radar goes away over the next year or two.
BTW I don’t see radar’s ability to see through fog as an advantage, because if the fog is that heavy your vision system won’t work anyway and you can’t drive on just radar alone.
Impact to TSLA:
Short term: none. The market is not smart enough to figure this out.
Longer term: Tesla widens autonomy lead against the current players dependent on active sensors (Waymo, Cruise) for localization and driving tasks.