"eventual projected" valuation probably means after the product is available on the market. Looks like right now it's all R&D and Testing.
Full level five autonomy without safety drivers is going to require extended R&D investments. However, with Google rolling around in cash the eventual outcome seems to be not that uncertain. Hence the insane valuation.
I'd say the main risk is increasingly other companies getting there first. E.g. Tesla seems to be quite stubbornly moving forward with its approach and unlike Waymo they have many vehicles on many roads. Also, the German car manufacturers are making progress.
Hard problems; but with feasible paths to a solution and a huge potential market once technology gets there. IMHO this won't take more than a decade and possibly a lot shorter. What will take longer is regulation catching up with reality. Safety drivers will likely be mandatory long after they stop adding any value whatsoever.
The reality is that most car accidents are caused by humans and increasingly in vehicles that don't have assisted & autonomous capable technology on board. Human drivers are by far the biggest risk factor on the road and autonomous driving is made harder by their presence. So, ironically, the job gets easier as that becomes rarer.
I’d consider that an extremely rosy outlook, and would argue nobody has yet demonstrated a “feasible path to solution” for level 5 autonomy (or in fact, lower levels). Self-driving seems as far away as it ever has been, much like fusion power.