In another thread we discussed the accuracy of predictions on the timelines of technology progress made by company leaders.
For example,
and
Let’s evaluate some current predictions, with outcomes not yet known.
Elon Musk, May 20 2025
Speaking on the number of robotaxis to be deployed over the coming months, starting with initial June 22 launch in Austin.
For the sake of specificity, let’s define “a few months” to be 3 months from June 22, say by Sep 30.
I am not privy to the proprietary info of AI driving companies like Tesla and Waymo and others, nor do I work in this area.
But looking at the current known robotaxi capabilities of Tesla (small geofenced area, small number of vehicles, safety monitors) and the historical pace of robotaxi rollout of Waymo (public rider-only started in 2020, fleet in 2025 is 1500 robotaxis), it seems to me that Musk’s prediction is too optimistic.
I’ll make the following counter prediction:
“By Sep 30 2025, Tesla will have deployed fewer than 1000 robotaxis that are actively providing robotaxi fares to the public in total over any/all US geographies.”
Please share other current predictions and any alternative predictions, for relevant companies and technologies.
Edit: Over any US geographies, not just Austin. Going to exclude international because me in US, me know even less international and China gonna do what China do.
Waymo blog, May 5 2025
A comment on adding to their robotaxi fleet in 2026.
If we interpret Waymo’s comment as:
“Waymo will add a net 2000 vehicles to its public, commercial robotaxi fleet by the end of 2026.”
then
I’d say I have no reason to not believe that prediction (but I’m open to available data that points either way).
Zoox, Amazon’s autonomous rideshare company, has just opened a massive new robotaxi production facility in Hayward, California.
At 220,000 square feet, the new production facility can make more than 10,000 vehicles a year. Zoox has said it will also create hundreds of jobs for people in the San Francisco Bay Area.
This is an easy prediction to make. It is almost a sure thing that they won’t have 1000 out there. Heck, it’s not even a sure thing that there will be ANY “actively providing robotaxi fares to the public” (by 30-Sep-25). It will take Tesla some time to analyze whatever data they are collecting in Austin, and then some more time to make whatever changes are necessary.
And if they remain only in Austin for a few months, they most likely wouldn’t need 1000 cars. And I think they want to use a different model primarily for their “real” robotaxi service, so they may be hesitant to use too many model Ys until that new model is ready.
If we, the moderately informed, think the original prediction is so absurd, why does the ceo keep making this kind of prediction, over and over and over again?
Because the CEO is “a character” and because everything he says “may be forward looking” and because he’s the official company cheerleader, but mainly because he’s a visionary, and visionaries very often have difficulty seeing the present in its reality!
I was trying to exclude having only employee riders and a fleet sitting in a parking lot not doing robotaxi stuff.
If Tesla has 1000 robotaxis doing what they are doing now in Austin (safety monitor, restricted public rider access), I count that as “robotaxi” business.
What do you mean by “restricted public rider access”? If you mean “by invitation” then they certainly wouldn’t need 1000 of them in Austin. Heck, even if they take over half of all the ride hailing business in Austin, they might not need 1000 of them!
or “access restricted to some subset of the public who are not Tesla employees”
See how helpful i can be by defining terms?
My prediction isn’t meant to be restricted to Austin, I don’t think they’ll have 1000 commercial robotaxis in total over any US geography. Again, like you say, not exactly a bold counter prediction.
That’s what they are doing TODAY!!! The people that were invited to try out robotaxi in Austin are not Tesla employees. I’m trying to figure out what additional subset of the public you are suggesting here that might bring the number of cars up to a higher number.
Ha! I don’t think they will have much of anything outside of Austin for a while. Possibly not until after September. They JUST STARTED collecting data in Austin, it’ll be a few weeks at best of collecting data, then all that data has to be analyzed, and then they can begin planning other cities. It’s not like there will be a magic switch to turn on robotaxi all over immediately. I mean, that’s the dream, but for now it will remain a dream. Maybe in a few years when there are lots of Tesla cars out there with the front bumper camera, and the latest hardware, etc they will be able to instantly turn a large set of cars into robotaxis. But for now I don’t think that is the case.
A thousand within a few months? No way! They have to do all the permitting, buy/lease/create the support facilities, hire a thousand safety drivers…ain’t happening.
LOLOL.
But, I do see your timeline.
You do NOT think there will be 1000 by Sept 30, 2025.
Ok. 3 months. Really… I’m thinking 3 months might not be enough time.
But, if it takes longer than 3 months … Then Tesla is not as far along with FSD n robotaxis, as I currently think they are.
Elon suggests 1000 in Austin, “in a few months”.
Ok. IF the “geofenced” area enlarges to cover the entire Austin metro area… Is 1000 doable? 4280 square miles (see below). I’m gonna say “yes, 1000 is doable”.
Especially if “in a few months” means 6 months.
Here’s my real prediction:
Elon gonna do what he thinks is reasonable when HE thinks it’s reasonable and doable.
And all the armchair CEOs will second guess - and intentionally FUD - the events.
ralph
I think FSD n robotaxis are Tesla’s market to lose - cause Tesla knows how to scale.
Waymo has had first mover advantage, but, as far as I can see, Waymo has NOT been able to truly seize that advantage. IMO, Waymo is still struggling with their technology.
ChatGPT says:
"The Austin–Round Rock–San Marcos Metropolitan Statistical Area (MSA)—the U.S. Office of Management and Budget definition for the “Austin metro”—covers roughly 4,279 square miles (about 11,080 km²) .
If you’re looking at a slightly broader regional view—like the Capital Area Council of Governments’ version that includes additional counties—the figure comes to around 4,300 square miles, which is even larger than the states of Rhode Island or Delaware combined".
That’s the problem. Work has expanded beyond the time available for its completion.
That didn’t matter back when Tesla was the undisputed leader in EVs and AVs. Musk could predict what ever he wanted. It didn’t matter. Tesla was still the leader. But now we’re to the point where not only Musk can’t deliver on what he promises, he can’t deliver what the competition is delivering either.
Remember, he not only promised to have thousands of robotaxis on the road by the end of the year, he promised they would be operating in California too. I don’t think it is too late, but the refrigerator door is starting to close.
I see Musk fired Tesla’s head of Sales and Manufacturing today. I suspect that will not fix the problem.
