Now that year-end sales figures are coming in for the Chinese market, it turns out that Xiaomi’s SU7 sedan has outsold Tesla’s Model 3 sedan in China. Which is interesting for several reasons, but one thing that stood out to me was that the SU7 is cheaper than the Model 3 and it comes with LIDAR ADAS standard:
The formula looks familiar if you’ve followed Xiaomi’s smartphone playbook:
Price: The base SU7 starts at RMB 215,900 ($31,000) vs. Model 3 at RMB 235,500 ($33,800), about 9% cheaper
Range: 700 km CLTC vs. 634 km for the Model 3 RWD, roughly 10% more
Tech: Free driver assistance features, LiDAR now standard on the updated model, and tight integration with Xiaomi’s HyperOS ecosystem
For a LIDAR unit to be included standard in a car priced lower than the Model 3 is a pretty significant indicator that mass-produced LIDAR isn’t anywhere near as expensive as folks think. Sure, it doesn’t mean it’s cheap. After all, the SU7 only has a single unit (compared to 4 in the current Waymo generation), and it’s not trying to power anything as strong as a truly autonomous unit. But it’s still a pretty strong sign that the costs of having LIDAR units drop significantly once you get into mass-production numbers…
While it isn’t cheap and it is a single unit, one might extrapolate that adding three more units would not substantially increase the price of the vehicle. One might hypothesize that you could get to four units total for another ~$10,000 or less.
I think that’s right. Once you get into mass-production, Lidar’s only adding a few pennies per mile onto the cost of a TaaS service. Waymo’s existing LIDAR equipped cars are super costly because they are hand-built. Well, hand-modified off a fairly costly base. But putting four LIDAR units into a production car as standard equipment probably just won’t be that big of a cost. If - or when - we get into an era where TaaS companies are in mass-market production runs of their vehicles, there probably won’t be that big a gap between the ones that use LIDAR and the ones that don’t.
Tesla really missed the boat in getting the hardware spec requirements wrong for their “ready for autonomy” cars under HW3 and (probably) HW4. The way to get a big advantage on vehicle cost in a TaaS system would be to use the huge amount of “sunk cost” people have in their existing privately-owned vehicles, already fully financed and heavily depreciated. Without that, they lose a big advantage over their competitors. If they end up having to use purpose-built vehicles the same way as everyone else, it’s hard to see them having much of a cost savings on the vehicle side.