Even at 50k revenue miles/year (100k total miles vs. 70k for Manhattan taxis) that's only 5 billion useful miles per year. Less than 0.2% of US VMT. That's a far cry from "significant value to a majority of Americans".
Furthermore, spreading 100k Waymos over 50 cities would be a colossal mistake, IMHO. (So that's exactly what Waymo will do, ha.... /s)
LA has ~100k Uber/Lyft drivers. Almost all work during surge pricing peaks, so you need 100k Robos to displace Uber/Lyft. 2k or even 5k Robos is a sub-scale, loser business model. It also provides no user benefit. You can't cut prices until you scale -- that's just Econ 101. Otherwise it'd be like Tesla cutting 3/Y to $25k tomorrow -- the only result would be ridiculously long waits and scalping, along with less cash flow to fund growth.
Waymo must first figure out how to displace Uber/Lyft in one city. They can then replicate that model in other cities. The resulting scale economies will allow them to think about chasing the big fish -- displacing 2nd/3rd car ownersihp in city+suburb metro areas. Thatwill "provide significant value to a majority of Americans".
What you are suggesting would be a terrible approach. It would likely mean spending billions and years just on one city. And there is no guarantee it would even work to displace Uber/Lyft. Waymo could have thousands of robotaxis that take up a big share but still not completely displace Uber/Lyft. And even if they did displace Uber/Lyft in one city, after years and billions of dollars, without robust and generalized FSD, they would not be able to replicate the model in other cities that easily since the autonomous driving would not work as well in the other city and require more testing and validation. So with your approach, they would spend billions and years just to scale one city at a time. That would not provide significant value and it would be a sure fire way to go under.
You seem to be forgetting that Waymo is not a robotaxi company. They don't need to displace Uber/Lyft. Waymo is a tech company, developing autonomous driving. Robotaxis are just one product. Ultimately, Waymo wants a portfolio of AV products including robotaxis, autonomous deliveries, trucking and even consumer cars some day. I believe ride-hailing is just a placeholder, a means to an end, because robotaxis are a good way of testing and deploying the tech while it is still in development. I believe they will also license their tech to consumer cars once the autonomous driving is sufficiently robust and generalized enough. Just focusing on robotaxis would be a mistake because it is thinking too small. If you just do robotaxis, yeah, you might get a decent share of the ride-hailing market but you will probably never displace Uber/Lyft completely. Focusing all your effort on displacing Uber/Lyft is shortsighted.
Waymo's stated goal is to develop robust, generalized autonomous driving that works in any city, on any vehicle. I believe that is the right goal because once you have that, you can do any AV product you want, from robotaxis, to autonomous trucks, to autonomous delivery, to self-driving consumer cars and they will work anywhere. You will get a much bigger share of the market and offer much broader value to the public than if you just do robotaxis.
And I believe Waymo has the right approach: use robotaxis to develop the tech, spread out to get your autonomous driving working robustly in diverse ODD towards that goal of generalized FSD. Once Waymo has robust, generalized autonomous driving, they can basically print money because they can do multiple AV business models anywhere and on any vehicle.
Waymo still hasn't figured out how to crawl.
Disagree. Crawling was Chandler. Waymo is way past crawling now. Waymo is jogging.