This may be a heavy hand to INTC (it was afterall a piece on INTC), but the overall picture on ADS was pretty good.
This is a pretty standard pic on the ADS levels
To get beyond today’s current best in class Level 3, many pieces need to fall together – some technical and some government approvals. The author states, quote:
“Intel will not be able to do this, as the roadshow presentation from CES 2020 also shows (see above). But this is not necessarily due to Intel. The competition is not ahead of Intel either. It is more that there is a need for multiple approvals and the infrastructure for enabling autonomous driving does not even exist yet. To make all this possible, a decisive development is needed and this development is 5G. 5G is generally considered to be decisive in this respect. However, this will change this year. With the coming rollout of 5G right ahead, I don’t think that the years 2024 and 2025, which Intel is aiming for, are unrealistic.”
Regardless of the hype, this is a long road (pun intended) … there are two things that will matter more than others: a) who has the largest and most ‘learnable’ dataset for AI / ML, and b) where will 5G infrastructure be reliable to send around the datasets required. (my opinion) … that completely skirts the government approvals (policy, etc).