Tesla plans to expand Robotaxi service to Bay Area.
Tesla plans to expand Robotaxi service soon
 |
| Hello, larger area in Austin. And possibly San Francisco. Credit: Bloomberg/Getty Images |
Tesla is looking to expand its Robotaxi service, and we're not just talking in Texas.
On Thursday, CEO Elon Musk said on X that the company plans to expand to "a larger service area in Austin this weekend."
Tesla's Robotaxi service launched last month in a geofenced area in South Austin, with a reported 10-20 Model Y cars, and it was only open to a small number of people. Musk didn't say whether the expansion also means introducing more cars to the mix.
However, the company seems eager to expand the service to other areas, too. Replying to a comment on X, Musk said that Robotaxi is coming to the Bay Area "probably in a month or two," following regulatory approvals.
The approvals shouldn't be too big of an issue given that Alphabet's Waymo (and, formerly, GM's Cruise) has been operating a self-driving taxi service for years in San Francisco and some of the nearby cities in the Bay Area.
SEE ALSO:Smooth, silent, strange: What it's really like to hail a robotaxi
Tesla's Robotaxi launch kicked off in June with multiple users praising how smooth the rides are, especially given that Tesla uses its regular Model Y cars for the service, as opposed to heavily modified cars used by Waymo and other self-driving taxi companies.
There were, however, many recorded instances of Tesla's robotaxis making odd errors, sudden stops, and apparent traffic violations.
The news comes shortly after Musk confirmed that his controversial Grok AI is coming to Tesla vehicles. The company also just announced its annual shareholders meeting will be held in November, a lot later than usual.
Topics Self-Driving Cars Tesla
Tesla will win and I'm not a fan boy
ReplyDeleteOh the horror! Elon's spreading.
ReplyDeleteI already have enough idiots on the road to watch out for. We don’t need this.
ReplyDeletestatistically speaking you are incorrect
Deletelol. There’s no statistical evidence to back up what you say.
DeleteIt's all hype. Musk got rid of lots of EV regulators but FSD without lidar is a joke. He has been promising it for ten years and the stock is still twenty times its real value.
ReplyDeleteThink long and hard about the young fella in California that watched his three friends burn and no way to get them out.
ReplyDeleteLOL
ReplyDeleteNot everyone wants to ride the bus sometimes having some peace and quiet on a ride is not so bad for the "everyday" people.
ReplyDeleteTesla now has a brand-new revenue stream that they didn't have before. They can scale it up however fast or slow they want. Their ability to scale is faster and cheaper than everyone else. They are definitely in the driver's seat.
ReplyDeleteSo it's going to fail in multiple cities?
ReplyDeleteYou can talk with Gork while your taxi blocks traffic in the rain.
ReplyDelete"Can I offer you an antisemitic rant in this trying time?"
DeleteBullish 🐂
ReplyDelete?url=https%3A%2F%2Fexternal-preview.redd.it%2Fw1qYYkJKhWfgmdW76flpmiVU8ON66TvcUgsL_HiE0Og.gif%3Fauto%3Dwebp%26s%3Dda1fc6f517caf93012ee6f146e3b77e2803a521d
DeleteYou mean I can get paid for this I’m willing to do it for free
ReplyDelete“3 months may be, 6 months definitely” vibe in this tweet
ReplyDeleteElon says in post on X "probably in a month or two
DeleteMore like "3 months maybe, 6 months also still maybe".
DeletePossibly, but he also said June for Austin back in January, and it did in fact turn out to be June for Austin.
DeletePeople keep saying this as though it were true. They said they would do an unsupervised service in June, and have yet to do one. A supervised service isn't remotely the same. It's not even 1 percent of what an unsupervised service is. It's like SpaceX promising you an orbital flight in June and they give you a suborbital one and thinking they met their target.
DeleteCrazy how I got banned from Tesla lounge and all the sister subs for saying this and mentioning how the Tesla hit the child sized dummy on a test.
DeleteThe delusion and separation from reality for them is fascinating
Yeah but who cares about that. We need to pump the stock! There is no time thinking about the real world.
DeleteFSD almost got us in a car accident yesterday, it ran a stop sign and cut of the oncoming traffic trying to make a left turn. Me and my friend were screaming in the car, he had to take over. Not sure who in their right mind would want to ride in a Robotaxi.
Deletethat was not robotaxi build software. Which has clearly better behaviors around emergency vehicles. Most likely school buses too
Deleteit looks like tesla tacked on the emergency vehicle behaviors. Likely school bus and school zone too
He also said nobody in the car for Austin, but there's somebody in the car.
DeleteI think the point was that they’ll probably find a way to keep their timeline for the Bay Area rollout even if it means a less automated system.
DeleteSan Francisco and California more generally will not be as liberal with permitting as Austin and Texas.
DeleteI’m not contesting that at all
DeleteSo far Austin has been a publicity stunt. Not a service. Not automated.
DeleteWait what? They announced taxis with drivers and chase cars back in January? I thought it was going to be a robotic vehicle without the need for human drivers and chase car.
DeleteRecently they have been hitting more of the advertised deadlines. Basically since Ashok took over.
Delete“3 months may be, 6 months definitely” vibe in this tweet
DeleteElon is a time optimist.
But do it really matter in the long run, if it take 12 months or 24 months?
Nah bro Elon said it will be coming in 2 seconds; when has Tesla ever delayed anything (sarcasm)
DeleteHave they even applied ? I don’t think they have.
ReplyDeleteI mean they have the drivered permits, but I don’t even think they have applied for driverless testing yet
They do not even have drivered permits. If they had drivered permits whether testing or deployment they would be PUBLICLY viewable on the CPUC site. They have a Chartered Carrier Permit. Tesla MAY HAVE APPLIED for one of the four possible permits administrated by the CPUC and DMV. None of the four applications have been approved or they would be available for public review. Driving your own employees somewhere is not a thing which is what they are doing in Palo Alto. California calls that a driver's license. Doing it in a company van requires a Chartered Carrier Permit. Exaggeration and fantasy are fun. Take them for what they are.
DeleteThe four permits in California are (1) Drivered Permit (2) Drivered Deployment (3) Driverless Permit (4) Driverless Deployment. Here's the real list or real companies with real permits. Save the link since this is live, it is real and does not involve unsubstantiated claims of what might happen in the future. When and if you get a permit you then must register each car including a VIN and provide real quarterly reports of operational details for public review. Once a permit exists the age of transparency begins at least in CA where the public gets a place at the table.
A reasonable analogy might be someday I plan to get a passport. I still have to pick up the form and fill it out, pay the fee and await its approval. If I don't have my passport by tomorrow it's another case of government interference I guess.
https://www.cpuc.ca.gov/regulatory-services/licensing/transportation-licensing-and-analysis-branch/autonomous-vehicle-programs/autonomous-vehicle-program-permits-issued
They've had the DMV permit for testing with a driver for a decade. They don't have a DMV driverless testing permit or a deployment permit.
DeleteLast I checked they didn't have any of the CPUC autonomous permits, either. They have the basic TPC permit that limo operators get.
Thank you. Seems odd
DeleteThe reason is they claimed FSD was an L2 feature with no plans to be L4 and that it was therefore exempt from California's required reporting program.
DeleteNow they're caught in a sort of a lie.
Thank you. Public Utility Commissions (PUCs) exist to protect the public via sensible regulation and reporting. I am glad there are rules to ensure that utilities calibrate my electric and gas meters and don't rip me off. Same for calibrating a gas pump. Seems insane to unleash autonomous driving on the public as a 'service' and not ensure we know about accidents in vehicles companies will CHARGE ME BY THE MILE TO USE. I would even be in favor of not allowing private judgments that hide settlement details from the public eye. Deregulation and even worse, intentional gutting of expertise in matters that are public facing just seems insane to me. I have years of experience with entities including the DOE, NRC, FAA and EPA to name a few. I believe you have to be an imbecile to think we'd be better off without such protections. I even know someone who was there at the onset when the CPSC was greatly reduced in scope in the 1980s in the name of efficiency. Even he shares with me now how the unintended consequences have been devastating.
DeleteDo you think they would attempt driverless testing already? They're not doing that in Austin yet, right?
DeleteYes but I believe if they were doing what they are doing in Austin in California.. they would need a driverless permit. Even if it’s not driverless
DeleteThey can just move the supervisor to the other seat. Putting them in the passenger seat was just for optics, basically just to make it clear they're not actually driving the cars and just supervising.
DeleteIf putting the supervisor in the driver's seat lets the cars start driving then they might just do it.
True, yea I can see them doing that
DeleteWaymo had driver in driver's seat for 20 months
DeleteSo just a more dangerous taxi?? Lol so stupid
DeleteHow do you know?
DeleteMuch of this information is public. It’s a tiered process where one permit relies on others that are public
DeleteSick good news coming out about Arizona application on June 26th.
DeleteWe will see what happens in California. My guess is they applied the same time.
https://imgur.com/a/d0kQU12
It is likely that they applied for one of the first stage California permits yes. That’s years away from driverless though
DeleteOkay dude
DeleteJust setting expectations
DeleteRight. I see that.
DeleteIt's not public yet. In the process in applying for DMV and CPUC.
DMV and CPUC process takes weeks to months. It's not published until approved or denied.
Right but there are prerequisites that will be public before applying
DeleteGiven they haven't reported any mileage or test results to California in years, I think the "regulatory approval" is going to carry a lot of weight on how long this really takes.
DeleteLol! They haven't even applied for the DMV permit for testing without a driver in CA.
ReplyDeletethey can't do that until they provide data WITH driver.
DeleteTesla has the ONLY permit they're "eligible" for.
https://www.dmv.ca.gov/portal/vehicle-industry-services/autonomous-vehicles/autonomous-vehicle-testing-permit-holders/
A manufacturer with a permit to test with a safety driver is authorized to test on any public road within the State of California. As of June 1, 2025, DMV has issued Autonomous Vehicle Testing Permits (with a driver) to the following entities:
Lol! Do they need one if they're going to testing with a driver in CA too?
DeleteWell then they would need to have the "safety monitor" in the driver's seat I believe.
DeleteYes. This is a requirement in California. It can’t be skipped there. They have the permit in place now to start testing this without paid passengers and with a safety monitor.
DeleteThey will also need to report all disengagements and accidents of course.
AFAIK, with a safety driver (not "passenger"), they already have a permit and could start right now.
DeleteYup
DeleteHere's what Tesla must do at each step of the process. The process is clear and comprehensive including public comment for some steps. There are four different permitting processes depending on what you are trying to accomplish. This document is just the 40 page overview. There are links for all the steps and all of the requirements and necessary reviews both internal and public along the way. In the last two quarterly calls, Elon & Ashok were clear. A public facing fully autonomous taxi service in Austin in June. They've made progress. Step one is not done as they are merely testing, it is not public and it is not autonomous as yet with a safety passenger doing live interventions. There are four paths offered in CA (Drivered Pilots, Drivered Deployments, Driverless Pilots and Driverless Deployments).I would assume Tesla will choose one of these and fill out an application and await approval from the city, county, CPUC and DMV in each case. If and when they get a permit they are online for public access including their detailed description of what they requested in their permit application. All of this is online so the public can just click rather than listening to unofficial claims from companies about what's happening.
ReplyDeletehttps://www.cpuc.ca.gov/-/media/cpuc-website/divisions/consumer-protection-and-enforcement-division/documents/tlab/av-programs/av-program-applications--general-guidance-august-2024.pdf
Great post, thanks for this.
Delete"The process is clear ... including public comment for some steps ... This document is just the 40 page overview ... all of the requirements and necessary reviews both internal and public ... await approval from the city, county, CPUC and DMV"
I'm going to say it's not clear but it certainly meets the "comprehensive" claim. This seems like a nightmare of a process, especially given it includes public comment periods, which are notorious for being time and labor-intensive and resulting is zero change to anything. I suspect that alone will tie the government up for months as they will get a flood of comments.
Really disappointed Tesla didn't pick a better 2nd service area like Atlanta. The process there is clear. You make sure you have 2x the minimum insurance and you start driving.
There are definitely pros and cons. Tesla has definite reasons they would favor California. Since they have bet on neural net convergence, there is no place on earth where they have more customer miles than the Bay Area. I remember even quite late in their history more than half of all the cars they EVER sold were in California. I did a bit of math a while ago that showed Tesla accumulates more miles in the Bay Area in a week or so than Waymo might in their lifetime. I did the math because an insider explained to me that for Waymo, real miles hardly matter at all since they have always used the TPUs to test synthetic miles nightly in GCP..
DeleteCA is challenging because they give the public a voice and make it hard for companies to bury the public right to know. Democracy is messy :) The flip side is Tesla has great data in CA and in the end for the taxi business case almost have of all the cities in the US with 5000 people / mi2 are in California. Save for Miami and its suburbs all of the rest of the cities in the American South are sprawling and not a single 5000 person/mi2 place. The ROI in taxis is density, affluence and tourism.
My WAG on the silly tweet from Elon is its just the "government oversight excuse" as a delay tactic. Tesla for $3K could have signed up for DRIVER TEST permit anytime in the last decade if they were serious. Complaining now about 'the heavy hand of government sounds like a 3 am riff from under a desk.
While I agree there are reasons Tesla choose CA, it's more to do with them having offices and teams there than anything else. There was an investment manager specializing in AV investment and they worked up a report that said Atlanta was ground zero for deployment of AVs based on their analysis. A large part of that is because Atlanta is the #1 major city in the US for miles driven per household at 36k compared to 22k for SF metro. There are also more households in Atlanta, with 2.5m vs 1.7m for SF. So a total addressable market of 90B miles vs 38B miles. Even more so since SF is really divided into two metros that are not going to be easily connected with AVs because of the bay so it complicates operations.
Delete"they give the public a voice and make it hard for companies to bury the public right to know"
Eh. Mostly it just stops progress. It's why nothing can get built in CA good or bad. Not saying other places don't have their issues, but I don't think CA has found a better one for sure. The public gets a say by electing leaders that will make the right choices for their city, county and state. Trying to let citizens have a voice in all the small stuff is destructive. There is a reason we're a Republic and not a direct Democracy. Now I'm 100% for regulations and publishing data openly. I'm 99% disappointed in what actually gets published though outside of NYC, which does a very good job.
"The ROI in taxis is density, affluence and tourism."
But taxis make their money on miles driven. I don't think density is a factor at all, other than you need a certain density to have enough demand. I also think raw population density is highly misleading. Where I live is 2500 people/mi2 but only 48% of the land in the city is residential which the rest being mostly offices and retail. So the residential land is 5000 people/mi2. Downtown Atlanta has Waymo in the core of Atlanta which is only 1,400 people/mi2.
Why affluence? AVs are expensive today, but they are obviously heading toward parity with cars or just a bit more. Pooled rides would be less. I see AVs as cost savings over owning a car for lots of people, especially those lower on the income scale that just need to supplement transit.
Tourists are certainly early low hanging frut, especially at $2/mile.
"Complaining now about 'the heavy hand of government"
Not sure if I consider what they are doing yet as complaining, but I don't disagree with your projection that they will when they miss the deadline they set for themselves.
Same approval they’ve never gotten in how many years??
ReplyDeleteFix FSD first
Wouldn't developing robotaxi fix FSD though?
DeleteEven if they need detailed mapping to truly go driverless, the other improvements to its decisionmaking should improve FSD performance everywhere.
You would think so, but we haven't had a substantive FSD update all year and their Robotaxi seems to be making the same kind of errors we've been experiencing intermittently on v13. I think Robotaxi has been more of a distraction.
DeleteIf one the necessary improvement is detailed mapping then it won’t do much to fix elsewhere.
DeleteAnd based on the Austin test it looks like there was more than mapping needed.
"looks like there was more than mapping needed."
DeleteRight, so any improvements to those would apply to FSD outside of robotaxi too.
Possibly.
DeleteYou never know what they will share between branches and how long it takes to do so.
For sure, especially for older vehicles on hw3
Delete1,000,000 robot taxis by 2019 vibes.
ReplyDeleteMy car is planning on traveling 300 miles pending stopping at the gas station
ReplyDeleteI am pretty sure it will do well in SF as I used it a lot there. But still having so many small problems in the east bay where I live.
ReplyDeleteI expect Arizona to open up first, even though the DMV says they only applied on June 26.
ReplyDeleteCalifornia is not Texas. You need to go through a rigorous testing program for all ADAS4 systems before they're allowed to commercialize. This is either a stock pump or an exposure that Tesla isn't adas4.
ReplyDeletePretty dumb “stock pump” if there is a chance at failing to get approval.
DeleteYou must be unfamiliar with this stock. Even bad news can fail to decrease the price.
DeleteRight but a “pump” implies they are doing something on purpose to inflate it. Makes no sense since insider sales never happen on these expected and normal milestones. A company taking prudent every-day steps isn’t a pump, it’s just progressing
Delete"...implies they are doing this on purpose"
DeleteTesla:
2016: robotaxis next year 2017: robotaxis next year 2018: robotaxis next year 2019: robotaxis next year 2020: robotaxis next year 2021: robotaxis next year 2022: robotaxis next year 2023: robotaxis next year
Yah their timelines are never on time lol. Everyone else who made the same promises were just quieter about it and none of their cars have delivered themselves.
DeleteSo there is no chance of Tesla successfully getting a drivered permit in CA and doing well?
DeleteCommercialized ADAS4? Not in a short time
DeleteHow much time do you think it will take? I assume this isn't a Tesla issue, if anyone wanted to get a permit it would take x amount of time and it's more than 2 months?
DeleteEnough to show that disengagements are rare enough to warrant approval.
DeleteIt took Waymo 20 months to expand its service area (details below).
ReplyDeleteIt is taking Tesla 3 weeks to expand its service area. It started opening up the service to additional customers beyond the first batch in under 2 weeks [https://teslanorth.com/2025/07/01/tesla-expands-robotaxi-access-in-austin/].
Details: Waymo launched its invite-only Early Rider Program in the Phoenix area in April 2017. The first expansion of service area occurred in December 2018 - it opened the service to a broader group of paying customers, beyond the original early riders and covered a roughly 100-square-mile area in and around Phoenix, including Chandler, Tempe, Mesa, and Gilbert. However it continued to use safety drivers behind the wheel for most rides at this stage.
No real reason to expand until you remove safety operator.
DeleteWhy not? More data.
DeleteIt's not as if the cost of the safety operator is of real concern for Tesla. Sure - it makes the service unprofitable. But until you go above 1000 cars, you're really just in early testing anyway.
"Pending regulatory approval" - heard this before, in 2019.
ReplyDeleteMight as well announce "Tesla Robotaxi expanding to Mars in 2 months pending successful colonisation."
ReplyDeleteWhen regulation Kills the inovation.
ReplyDeleteBetter than the "innovation" killing a person.
DeleteMusk is forcing this whole Robotaxi thing through even though it’s far from ready to do full driverless. It’s so obvious.
ReplyDeleteHow is it so obvious?
DeleteHe's so cringe, trying desperately to pump the stock.
DeleteIf Tesla cared about a working product, they would focus on removing the safety driver
They are focusing on removing the driver, it should not be rushed though
DeleteThey can do both. In CA you can't start without a safety driver so they can get started on that.
DeleteEdit: Tesla will be illegally using FSD with drivers in sf
ReplyDeleteLol they haven't figured out "FSD" yet. They want a baby dick foot print wherever waymo is
ReplyDeleteNot sure why the downvotes. This is accurate. They are just trying to create the perception of expansion. But until they remove operators in cars they can’t expand by a few cars in each GEO.
Delete