Sponsored

Gen(R3)Xer

Well-Known Member
Joined
Mar 10, 2024
Threads
46
Messages
1,389
Reaction score
1,354
Location
Ohio
Vehicles
Leasing Model 3 until R3X comes out, but now I have an R2 reservation as well.
Rivian has me ordering in the Sept-Oct time frame. Seriously wondering if I should just wait a few months more for Lidar? Tesla and a Chinese EV company have said no to Lidar, so I'm wondering what the best move is?
It’s not just LiDar, but a more powerful in-house chipset (RAP1) that somewhat future-proofs your purchase for years to come.

The current NVidia chip set, while powerful and capable of Level 2+ autonomy, will never be able to achieve Level 3 or above. It will always have to be supervised, which is fine for many people.

The hope is that the LiDar, and the more powerful RAP1 chipset, will accelerate Rivian’s attempts at autonomous driving and help them to catch up to Tesla. They don’t have the volume of data that Tesla has (millions of vehicles and billions of miles), so they have their work cut out for them.

I’m excited to see what they can do. I’m not a huge autonomous driving fan myself, but I recognize that there’s a growing interest in its capabilities. If they “crack the code” I’d imagine it would increase the resale value of the vehicle as well. We’ll find out in the next few years.
Sponsored

 
Last edited:

R1SFamilyGuy

Well-Known Member
Joined
May 3, 2023
Threads
28
Messages
365
Reaction score
281
Location
NJ
Vehicles
Mazda CX-5
Occupation
I do software things
Noob question. Maybe something applicable to me and a small set of users.

Given that I look at self driving as a convenience feature not something that let’s me sleep at the wheel etc, I do not think Lidar matters to me. Tricky lighting, weird roads, aggressive traffic I want to be alert and be ready or already taken over the driving. While lidar may be needed for robo taxis, I really don’t. I think.

Am I wrong or missing something?
 

sparked

Well-Known Member
Joined
Dec 16, 2025
Threads
12
Messages
245
Reaction score
389
Location
USA
Vehicles
R2 Deposit
Noob question. Maybe something applicable to me and a small set of users.

Given that I look at self driving as a convenience feature not something that let’s me sleep at the wheel etc, I do not think Lidar matters to me. Tricky lighting, weird roads, aggressive traffic I want to be alert and be ready or already taken over the driving. While lidar may be needed for robo taxis, I really don’t. I think.

Am I wrong or missing something?
RJ thinks that autonomy is going to move faster than people think over the next few years. Nobody is going to argue with your cautious viewpoint though. It's a valid take to be skeptical and cautious with autonomy. No right or wrong here as we just have to see how this plays out.

I think we will understand Rivian's progress a little better at the end of the year when they release Point to Point.
 

r2fb

Well-Known Member
Joined
Apr 16, 2026
Threads
0
Messages
60
Reaction score
67
Location
NJ
Vehicles
Tesla Model Y, Audi Q4
Tesla also went vision only and continues to lead in autonomy, so while I'm sure LIDAR will help push Rivian forward, its clearly not required to match Tesla. I do wonder if part of Tesla's decision to go vision only was because they already had so much training data from their LIDAR-equipped models and didn't need more.
Tesla leads in supervised autonomy which is the best they can do with cameras only. Waymo leads in unsupervised autonomy which is the real prize given the FSD take rates in Tesla.
 

Sponsored

shandering

Well-Known Member
First Name
Shane
Joined
May 26, 2026
Threads
0
Messages
96
Reaction score
39
Location
Atlanta
Vehicles
Polestar 2
Occupation
Artist
Tesla leads in supervised autonomy which is the best they can do with cameras only. Waymo leads in unsupervised autonomy which is the real prize given the FSD take rates in Tesla.
Everybody gets this wrong. E2E neural networks seem to use camera as a backbone. There is no evidence that any chinese car with lidar can avoid any accident that tesla can't with cameras only.

in fact it's always been the opposite.

People also claim that compute will give rivian an edge, but I doubt rivian is even years from matching the performance of FSD v14 lite on HW3

The question is whether you can drive with end to end neural networks and no HD maps. Only tesla is even close to demonstrating a safety level where this is possible.
 

Eric9610

Well-Known Member
Joined
Nov 27, 2023
Threads
16
Messages
792
Reaction score
876
Location
So. Cal.
Vehicles
R1T QM, R1S QM, MB E350, MB S63C
Clubs
 
I don't think the inclusion of lidar sensor is the bottle neck. I don't think we see any lidar cars till mid to end of 2027 for ordering. The bottle neck is the Rap 1. Knowing how constrained chips are and production spots are limited this will be the holdup.

Anyone getting a launch package car will not have lidar no matter when ordered.
 

r2fb

Well-Known Member
Joined
Apr 16, 2026
Threads
0
Messages
60
Reaction score
67
Location
NJ
Vehicles
Tesla Model Y, Audi Q4
The question is whether you can drive with end to end neural networks and no HD maps. Only tesla is even close to demonstrating a safety level where this is possible.
All of the "dozens" of unsupervised Tesla robotaxis use much higher quality maps for the geofenced regions, manually marked and curated based on driving data. Everything else is supervised, and supervising is driving. It's just an assist, nothing game changing as evidenced by its lack of popularity. You either have a cab that drives for you and you don't need to pay attention, or a good driver assist system and no driver assistance is going to transform the way we travel.
 

rfkxyz

Well-Known Member
Joined
Mar 25, 2026
Threads
1
Messages
170
Reaction score
153
Location
Virginia
Vehicles
2004 XC90, 2024 Grand Highlander Hybrid
I wonder if we might see an early flip to MY28? Had assumed summer 2027 based on R2 history. Maybe ~Jan 2027?
They won't ship LIDAR units until LE trim is over IMO. Can't see them giving you LIDAR/RAP1 and LE package, which includes Autonomy+.
I think one of the Rivian engineers said LIDAR/RAP1 would be in the "next model year" which means 2027 debut. I agree that it is unlikely to overlap with LE, though I do wonder if we will see non-LIDAR Premium by the end of this year while the Performance is still wrapping up LE?
Perhaps MY27 and LE run concurrently & exclusively until early-2027 brings LiDAR and ends LE for early flip to MY28?

Hard to say if offering Premium alongside LE makes much sense or wait for LE to die off?
 

Jeremy3292

Well-Known Member
First Name
Jeremy
Joined
Apr 27, 2026
Threads
5
Messages
587
Reaction score
877
Location
South Carolina
Vehicles
R2
I've actually changed my mind on this. I was going to wait for LIDAR/RAP1 but there's too many unknowns as it is. LE trim comes with Autonomy+ included. They promised point to point by EOY on the current chip set. That's where Tesla is at now - not eyes off L3 - which will require RAP1/LIDAR. So if Rivian gets as good as Tesla is today on the current chipset then perfect, no worries from me. I wasn't going to pay $2,500 for it even if I got a RAP1/LIDAR vehicle probably. And who's to say they won't increase the price next year for it? Seems one in the hand is worth two in the bush. Just give me good point to point like Tesla, that's all I personally need.
 

Sponsored

tivoboy

Well-Known Member
First Name
Paul
Joined
Nov 8, 2025
Threads
2
Messages
266
Reaction score
272
Location
Palo Alto, CA
Vehicles
M3
Noob question. Maybe something applicable to me and a small set of users.

Given that I look at self driving as a convenience feature not something that let’s me sleep at the wheel etc, I do not think Lidar matters to me. Tricky lighting, weird roads, aggressive traffic I want to be alert and be ready or already taken over the driving. While lidar may be needed for robo taxis, I really don’t. I think.

Am I wrong or missing something?
You’re not missing anything really, although nobody really knows EXACTLY what is going to transpire.. But if one is NOT looking for full hands off / eyes off autonomy, or robotaxi type of capability (there is a bit of confusion on whether or not Autonomy+ WILL indeed allow much more hands off eyes off than CURRENT autonomy, but not robotaxi type L4/5 capability) you’re only going to miss out on being in the bleeding edge, possibly is 2-3 years (I doubt Rivian is going to “beta” future L3 Autonomy across their fleet anytime in 2027 or even possibly 2028.

And you MIGHT miss out in the context of possibly a bit more accelerated depreciation of a current NON-LIDAR vehicle, going into H2’27, but I doubt it’s going to be more than an additional 5-7%, probably not 10%,,

The value proposition for the current LE I think for this use case, really does overweight now vs. wait certainly for a usability standpoint.
 

shandering

Well-Known Member
First Name
Shane
Joined
May 26, 2026
Threads
0
Messages
96
Reaction score
39
Location
Atlanta
Vehicles
Polestar 2
Occupation
Artist
All of the "dozens" of unsupervised Tesla robotaxis use much higher quality maps for the geofenced regions
that is false. They use the same level of maps found on consumer cars. This is not HD maps in the sense that HD maps are lidar-scanned, centimeter accurate maps that are synchronized to the lidar. By doing this you have a centimeter accurate positioning of your car within the world and can path plan and drive with no lane markings or camera visibility.

There is no evidence this is easy to do with end to end transformer neural networks. And any cars lacking 3d spinning lidar (or at least 360 lidar coverage) will not be able to do this anyways

Tesla uses consumer level car GPS which has very low accuracy and can be 15m off from where the position of the map is. You need very advanced driving to get around this because you no longer can rely on the map for accuracy

By having centimeter accurate maps (ala waymo) you can map speed bumps, potholes, construction zones, geofence very specific areas the car can't go, put invisible walls the car can't drive through, etc.

The basic level of driving required to do unsupervised is very low with HD maps. As you reduce reliance on the maps, you get waymo of today where they make more mistakes and the driving is worse. The hard part is driving around in parking lots (which need more advanced knowledge and people/cars are going everywhere) and pickup/dropoff. Two things which zoox does very poorly. They don't even drive in parking lots.

manually marked and curated based on driving data.
This is also false. The reuters report was about data labelers. They have nothing to do with mapping. Tesla was training robotaxi specific features like pickup an dropoff while avoiding bike lanes. Tesla was using austin data to do it since they had to do test driving anyways.
 
Last edited:

Megaman0025

Well-Known Member
First Name
Justin
Joined
Jun 8, 2023
Threads
0
Messages
102
Reaction score
82
Location
16066
Vehicles
Tesla Model 3 Performance
Does make one wonder if this gives hope to "late 2026" for Gen3 Lidar inclusion in builds... I mean, if the hardware is fully validated, even if not turned on, might they start including it soon. Who knows.
Honestly, waiting for the RAP 1 chip is better move than worrying about Lidar. That chip is what is going to enable the much improved ADAS and self driving compute needed in the future.
 

savethemanual

Well-Known Member
Joined
May 18, 2024
Threads
17
Messages
953
Reaction score
1,630
Location
Earth
Vehicles
Future R2
Honestly, waiting for the RAP 1 chip is better move than worrying about Lidar. That chip is what is going to enable the much improved ADAS and self driving compute needed in the future.
Until RAP2 comes out, then RAP3 and so forth....there's always going to be the latest and greatest. At some point you just have to jump in. I'll be coming from an ICE with only blind spot monitors 🤣 , so getting the LE R2 will be a massive upgrade and I'm smitten with that!
 
 








Top