Texas Is Leaving San Antonio Behind on Robotaxis

I Texas Is Leaving San Antonio Behind on Robotaxis

I went looking for when City Council was going to vote on the Tesla lot.

Back in July the Zoning Commission recommended approving it, eight to two. A vacant parcel on West Josephine, a few minutes off Pearl, next door to a pre-K through eighth grade school, rezoned so Tesla could lease it and run a private charging yard for 56 robotaxis. City staff had recommended denial. It made the papers for about a day.

So I pulled up the agendas to find the real vote.

The case is Z-2026-10700135. It was Item 7 on the August 20 Zoning and Land Use session. If you watch that meeting, about a minute in, the clerk says items 7, 13, 22 and 23 have been withdrawn and will not be considered today.

That’s the whole record. It isn’t on the September 3 agenda either. I couldn’t find a single news story about it and I looked hard.

One thing I do not know is what withdrawn means here. Maybe the applicant pulled it. Maybe it’s getting renoticed. Maybe there’s a negotiation going on that nobody’s written up. I’d rather tell you I don’t know than guess, and anybody telling you confidently what happened in that room is guessing too.

What I do know is that six weeks after a city panel voted to approve it, the item came off the agenda and nobody in this city said a word about it.

That’s the post. The rest of Texas is scaling this technology right now, and we’re the one big metro still arguing about a parking lot.

San Antonio doesn’t say no to things. We take so long to say yes that the decision gets made somewhere else, and then we act surprised about where it landed.

We’ve done this before.

Where the robotaxi thing actually stands

Most people here have half this picture, so let me put the whole thing down.

Tesla runs commercial robotaxi service in Austin, Dallas and Houston. Austin started in June 2025 with a safety monitor in the passenger seat, dropped the monitor in stages, and went unsupervised across the metro by June of this year. Dallas and Houston came online April 18. Miami, Orlando and Tampa followed. San Antonio is not on that list, there’s no announced date, and the only piece of San Antonio infrastructure Tesla is publicly known to want is the lot that just disappeared off the agenda.

Waymo is here and has been since February 24, when it opened to invited riders in the same batch as Dallas, Houston and Orlando. Airport service on March 31. Roughly 60 square miles.

Then April. On the 4th, an empty Waymo got into trouble with high water near McCullough and Contour. On the 20th, during heavy rain, another empty Waymo was swept into Salado Creek and turned up four days later downstream near the Greenway trail. Nobody was in either car and nobody got hurt.

Waymo pulled rider service here and filed a recall with NHTSA. Recall 26E026, filed May 6, covering 3,791 vehicles, which is the entire fifth and sixth generation fleet in every city the company operates. The defect language in the filing is worth reading. On higher speed roads, the vehicle “may slow but not stop” when it detects a flooded lane. San Antonio rain wrote a software patch for Phoenix, San Francisco and Los Angeles.

Service still hasn’t come back for riders here. Waymo said in May it planned to return, and the company’s own status page lists San Antonio as a city it serves, but you can’t hail one. The pause here was already being called the longest the company has had in any city back in April.

Now the part nobody wrote up locally. On May 13, Waymo published a post about growing to 1,400 square miles across 11 cities. The expansions were Miami, Austin, Atlanta, Houston and the Bay Area. Not us. And when the company dropped its waitlist and opened service to everybody, it logged Dallas on August 4 and Houston on August 20 in its own announcement log. There’s no San Antonio entry in that log at all.

So the one operator that did show up launched us in a batch of four, then expanded the other three, opened the gates in the other three, and still hasn’t turned us back on.

We aren’t being ignored by the industry. We’re being deprioritized inside it, by the company that already showed up.

One more number, and I want to be precise because this one gets repeated wrong. Texas DMV made its automated vehicle database public on May 28 under new state rules, and Bloomberg was first to report what was in it. Waymo had registered 577 automated vehicles in Texas. Tesla had 42. Those are vehicles authorized with the state, not cars out driving. Independent trackers had Tesla actually running about 20 in Austin, three in Dallas and six in Houston at the time. Use it for the ratio, not the roster, and know it’s three months stale.

Why Austin, Dallas and Houston got it and we didn’t

This is my opinion, not a citation.

Austin is obvious. Tesla’s headquarters and its factory are there. The engineers who fix the cars live there. When something goes sideways at 11pm, somebody can be standing next to the vehicle in 20 minutes. You launch the hardest version of your product in the city where the people who built it sleep. Any of us would do that.

Dallas and Houston are the ones San Antonio should be studying. They’re bigger and denser, which matters because trips per square mile is the number that decides whether a robotaxi pencils out. But they were also easy. Nobody in Houston made a company spend a summer at a zoning podium to get a place to plug cars in.

And that’s where I think people here have it backwards.

The state already took away our ability to say no. SB 2205 in 2017 and SB 2807 in 2025 put automated vehicles under exclusive state control, and Section 545.452 says a political subdivision may not impose a franchise or other regulation on the operation of an automated vehicle. San Antonio cannot ban robotaxis, cap them, license them or write rules for them. That’s settled.

What we still control is dirt. Zoning, land use, permits, the physical places these fleets have to live. It’s the only lever we have left and it’s the one we’re slow with.

I’m not saying the neighbors were wrong. A commercial car yard next to a pre-K through eighth grade school is a fair thing to argue about, and Commissioner Loseff wasn’t being unreasonable when she said the corridor is supposed to be walkable and dense. Staff recommended denial for real planning reasons. The applicant’s attorney had real answers about power capacity and routing the cars off the school frontage. Two defensible positions, nobody acting in bad faith.

What I’m saying is we have no way to resolve that in under a year. We have no pre-zoned sites for this and no policy for where fleet depots go. Our own Zoning Commission chair said out loud in July that the development code needs updating for autonomous vehicles, which is another way of saying we’re making it up case by case while three other Texas cities already run the service.

An operator looks at that and doesn’t get mad. They go somewhere else first and come back to you in year three.

We did this to Lyft

If this feels familiar, it should.

Uber and Lyft both launched here in March 2014. The police chief held a press conference, issued a cease and desist, and SAPD started writing citations at up to $500 apiece. That December, Council passed a rideshare ordinance seven to two: fingerprint checks through SAPD, drug testing, inspections, permit fees, two weeks of processing in person. Rey Saldaña and Ron Nirenberg were the two no votes.

On April 1, 2015 both companies stopped operating inside city limits. Uber then signed agreements with Windcrest, Alamo Heights, Olmos Park and Hollywood Park and kept running in the suburbs while it was banned in the seventh-largest city in America.

By August the city had folded. Council approved a pilot, six to five, that made the fingerprinting voluntary, which gave up the one thing the entire fight had been about. Uber came back that October, Lyft that December. Then in May 2017 the governor signed HB 100, which created a statewide framework, expressly did not require fingerprinting, and erased every local rideshare ordinance in Texas including ours.

Roughly 32 months of fighting, losing the service and renegotiating a local rule that the Legislature wiped out six months after we finally settled it.

Marina Gavito, who ran Tech Bloc, said the line I still think about:

“When Rackspace is trying to hire a senior vice president moving from Chicago and she arrives at the airport and pulls up the Uber app and it says Uber is not available… That’s instant brand damage to San Antonio.”

Notice where that quote starts. The airport.

The airport, and why you can’t find a lounge

Fly out of SAT and look for a lounge.

There’s a United Club in Terminal B, and it’s only there because Continental built it before the merger. There’s a USO for military and dependents. That’s the list. No Admirals Club, no Sky Club, no Centurion, not one Priority Pass lounge in the building. Seventh-largest city in the country.

The fair version, because I’d rather check the boring explanation before I get mad about something. Terminal C has more than 29,000 square feet of club space in the budget, funded and drawn. What it doesn’t have is a single airline that has announced it’s moving in. American has been reported to be building an Admirals Club here and has never put out a release saying so. Delta’s CEO has mentioned San Antonio on a podcast. Meanwhile American formally announced a new Austin club in March, and Delta already has a 30,000 square foot Sky Club there.

Same pattern as the cars. The building goes up, the brands show up later, after they’ve done the other city first.

How we got Terminal C is its own version of this. A committee to study the airport’s future in 2017. Consultants delivered in 2018. Council approved in November 2021. Broke ground December 2024. Terminal C opens in 2028. That’s 11 years from “let’s look into it” to a gate you can board from, and the 1994 master plan had already called for new terminals that never got built.

Credit where it’s due, because I’ve been told my last few posts were short on it. Terminal C is on schedule, on budget, and slightly ahead of what the 2021 plan promised. The construction is going fine. All the slowness is at the front end, in the seven years we spent deciding.

Then there’s Southwest. Our largest carrier got assigned gates in Terminal A, which our own documents call functionally obsolete and undersized, while the new terminal filled up with everybody else. Southwest wanted 10 gates in Terminal C. It refused to sign a 15-year lease, then sued the city in September 2024. The city’s filing called it an extreme case of sour grapes. Southwest lost in federal court, appealed, filed an FAA complaint and asked the federal government to withhold grant money from our airport. The city stripped Southwest’s branding off its gates and stopped cashing its checks.

It settled on May 14 of this year, 20 months in. Southwest ended up with three gates in Terminal C and three in a renovated Terminal B.

Now put that next to Austin. On January 7, Austin announced that Southwest, Delta, United, American, Alaska, FedEx and UPS had all signed 10-year agreements at once and called it historic. It was a press release.

Southwest is getting 18 gates in Austin’s new concourse. It’s getting three in ours.

We got there through a lawsuit, an FAA complaint, federal mediation and pulling the logos off our biggest airline’s gates. Austin got there with a photo.

And last year SAT lost passengers for the first time since 2020, down about 3% to 10.7 million, while Austin sat at 21.7 million.

The pattern

Rideshare. The airport. Robotaxis. Same shape every time.

We don’t reject the thing, and I’m tired of anti-technology being the framing, because it isn’t true. What we do is take the maximum available amount of time to reach roughly the answer everybody else already reached, and then wonder why the next company starts somewhere else.

And I want to be direct about what’s happening now, because it looks different than 2014 and it isn’t. San Antonio is fighting robotaxis. We just can’t fight them the way we fought Lyft, because the state took the ordinance away from us.

So the fight moved. It’s zoning now. It’s a conditional use permit for a taxi parking lot. It’s a development code that has no category for this and a staff recommendation of denial because a car yard doesn’t match the walkable corridor plan. It’s an item coming off a Council agenda in August with no explanation. None of that is a vote against robotaxis, and all of it works like one.

That’s the version of a no that a city still has, and we’re using it without ever having to say it out loud or take a position anybody could argue with. In 2014 we at least told Uber and Lyft what we wanted. Right now Tesla’s landlord doesn’t know what we want, because we’ve never decided.

The cost isn’t the technology. It’s that being late means somebody else set the terms. Somebody else drew the service map, decided whether the flood-prone side of town gets covered, decided whether your biggest airline gets three gates or 18. Whoever’s early writes the rules. Whoever’s late hosts the leftovers and argues about where to park them.

How we get in front of it instead

The good news is that none of this needs permission from the Legislature. Every lever here is zoning, data or a utility timeline, which are the things a city still controls. Five moves, and only two of them cost money.

1. Pre-zone the depot sites now. Pick six or eight parcels near Port San Antonio, the airport and the industrial ring where a fleet yard is allowed by right, and publish the list. The next operator shouldn’t have to discover our zoning process the way Tesla’s landlord just did.
2. Update the development code for autonomous vehicles. Our own Zoning Commission chair asked for this in July. Write it down before the next application shows up and we relitigate the whole thing block by block.
3. Say what happened to Item 7. Not a conspiracy, just basic disclosure. When a case with this much public interest comes off the agenda, the city should be able to say in one sentence whether it’s dead, deferred or being renegotiated.
4. Publish the flood data as a product. Low-water crossings, closure history, drainage projects, live alerts, in a format software can read without anyone making a phone call. We’re the city that produced a national recall. Be the city with the best flood data because of it.
5. Put CPS Energy in the room on day one with a published interconnection timeline. The attorney on the Josephine case said Tesla picked that corner for the power. If our answer to “how fast can I get power” is a phone tree and somebody else’s answer is a webpage, we lose before anyone files anything.

Do those five and San Antonio stops being the metro where a fleet has to fight for a place to park. That’s the whole ask. We’re not going to win the first launch of anything, and I’ve stopped wanting us to.

We can be the easiest place in Texas to run the second one. Right now we’re the hardest, and we did that to ourselves.