I think you missed the technology of flash Lidar, which is much cheaper and more powerful than traditional Lidar. It is most commonly implemented using low cost TOF cameras, which we already see ubiquitous in mobile phones, especially those from Huawei.

That system, coupled with machine intelligence in the graphics processing hardware, enables a single moving phone to track all of the moving joints in multiple walking humans (“Motion Capture Skeletal tracking”), or any kind of animals, in real time, as well as the static terrain of course.

Presumably this would be the camera technology used by Tesla, as well as probably all of the Chinese self driving car manufacturers.