Tesla's Robotaxi service in Austin, Texas grew again on November 9, 2026: the company's official Robotaxi account posted a map showing a wider geofence, and independent tracker Robotracker.app logged the active area at 435 square miles, up from 272 mi² on October 23. That is roughly a 60% area increase in two weeks — not the 443 square miles some wires copied from an older 443 km² milestone.
Readers care because geofence area is the public scoreboard for unsupervised robotaxi rollouts — alongside fleet counts, hours, and safety cases. Tesla is racing Waymo and its own multi-city launches (Fort Worth, San Antonio) while regulators and councils debate caps elsewhere, as in our DC Waymo bill coverage.
What changed on November 9
Autonomous vehicle sensing arc over a small car in an expanded service zone
EVwire's November 9 brief says Tesla's Robotaxi account shared old versus new boundaries, expanding in most directions except north — toward Manor, Barton Creek, Manchaca, and past SH-130 in EVwire's map read. They estimate ~440 mi² (about 1,140 km²), 62% larger than the 272 mi² zone it replaced.
Robotracker — which scrapes operator maps — records the same event as 272 mi² → 435 mi² on its updates feed. Small differences between human map scaling and automated geofence extraction are normal; both point to low four hundreds of square miles, not 170.
The 443 number is usually kilometers, not miles
Confusion comes from history:
| Date (tracker log) | Area | Notes |
|---|---|---|
| Aug 26, 2025 | 443 km² | ~171 mi² expansion logged by Marco Gonçalves |
| Oct 23, 2026 | 272 mi² | Pre-November 9 footprint |
| Nov 9, 2026 | 435 mi² | Current Robotracker reading |
Not a Tesla App's table mixes km² and mi² rows; copying 443 without units produced today's bad 443 mi² headlines. For editorial accuracy, cite mi² from Robotracker or EVwire's mile estimate, and label km² when quoting 2025 archives.
Austin versus Waymo on the same tracker
Sensor halo around a vehicle comparing operator coverage footprints in one metro
On November 9, 2026, Robotracker's Austin city page listed:
- Tesla: 435 mi²
- Waymo: 150 mi²
Area is not rides, crashes, or DMV permits. Waymo's fleet scale targets and Toronto 2028 plans show Waymo playing a multi-city regulatory game while Tesla optimizes for rapid geofence growth in Texas.
Texas DMV data cited by EVwire put 1,282 vehicles registered to Tesla's Robotaxi service in Texas as of late October, with Cybercab models representing a large slice after public Cybercab rides began September 4. Operating hours reportedly run 5 a.m.–2 a.m. with Tesla aiming for 24/7 by end of 2026 on its earnings call — verify in app before planning trips.
Timeline: from 20 mi² launch to metro-wide Austin
Austin has been Tesla's unsupervised robotaxi laboratory since June 22, 2025 with a ~20 mi² launch footprint. Robotracker's ladder shows steady growth through 2025 testing, September 2026 public app activation, and October Houston/F1 expansions.
November's jump matters because it approaches metropolitan scale inside one US Sun Belt labor market — relevant to Uber's driverless volume forecasts and insurer models, even if Tesla does not publish ride counts.
What riders and regulators should verify
Before treating expansion as full-city autonomy:
- App pickup pins — polygons can lag marketing maps by neighborhoods.
- Vehicle mix — Model Y fleet versus Cybercab two-seaters may differ by zone until Tesla confirms parity.
- Airport access — earlier expansions added Austin-Bergstrom; confirm airport eligibility in-app after this polygon change.
- Safety reporting — Texas disclosures and NHTSA standing general order data still underpin trust more than square miles.
Policy readers should cross-read Volvo-NVIDIA HALOS door-to-door claims: highway consumer autonomy and geofenced robotaxi expansions are different products on different clocks.
Why the unit mix-up matters for AI coverage
Robotaxi stories travel fast on social graphs with single numbers. Mixing km² and mi² overstates Tesla's lead by nearly 2.6×. When you cite expansion, name the tracker, date, and unit — same discipline we use on claimed-vs-verified GPU lease stories.
For builders working on perception or fleet sims, geofence diffs are a free dataset: diff Robotracker snapshots when municipalities negotiate VMT fees or vehicle caps.
Texas fleet scale and operating window
EVwire ties geofence growth to fleet counters, not just map paint. Texas DMV records cited there showed 1,282 vehicles registered to Tesla's Robotaxi program in late October, with Cybercabs making up a large fraction after public Cybercab rides began September 4, 2026. Hours reportedly span 5 a.m. through 2 a.m., with Tesla telling investors it wants 24/7 service by year-end 2026 — verify in the app because maintenance windows still appear in some cities during early rollouts.
Compare to Uber's public driverless targets: Tesla is optimizing owned fleet geography first, while ride-hail partners optimize utilization per vehicle-hour. Analysts should not multiply square miles by fleet count to guess revenue — utilization and safety interventions dominate.
Regulatory and liability context (high level)
Texas has been permissive relative to California on autonomous testing and paid rides, which is why Austin became Tesla's unsupervised proving ground in June 2025. That does not remove NHTSA reporting, insurance, or municipal negotiation when cities impose fees. Austin itself has not mirrored DC's vehicle cap drama in our coverage set, but statewide Democratic gains discussed in The Register's midterm piece may eventually surface in transport committees.
If you build maps or compliance tools, store polygon versions with timestamps. Regulators comparing operators will ask what was live on date X, not what a blog remembered.
Cybercab versus Model Y in the same app
Tesla's November 9 post did not promise Cybercab-only service for the entire new polygon. EVwire notes 573 Cybercabs in Texas registrations in one snapshot — substantial, but not the whole 1,282 figure. Riders may still receive Model Y units depending on supply.
Capability differences matter for accessibility (Cybercab's two-seat layout) and sensor suites. Third-party testers should log VIN class per ride when benchmarking disengagements.
Competitive dynamics with Waymo in the same metro
Robotracker's 150 mi² Waymo versus 435 mi² Tesla snapshot is operator-reported geofence, not passenger market share. Waymo's business leans on regulatory relationships in California, Arizona, and expanding cities like Toronto 2028 plans. Tesla's strategy is frequent boundary pushes in Texas plus rapid multi-city ignition (Fort Worth 370 km² on November 3 per Marco Gonçalves' log).
Neither metric substitutes for crash and intervention rate transparency. Readers evaluating AV stocks should insist on disengagement and injury disclosures, not map acreage alone.
Historical expansion ladder (why November feels fast)
Marco Gonçalves' public log shows Austin alone absorbing dozen-plus polygon revisions since June 22, 2025. October 2026 brought F1 Circuit of the Americas service tweaks and a Houston expansion an order of magnitude larger than its launch footprint — context for how aggressively Tesla iterates maps when software confidence rises.
November 9 snaps the metro geography from ~272 mi² to ~435 mi² in Robotracker — a step change, not a marginal buffer. EVwire notes San Antonio had briefly held largest single-city geofence honors earlier the same day before Austin retook the title. That intra-brand competition is unusual: most operators optimize cities sequentially; Tesla's public narrative rewards headline races between Texas metros.
For simulation engineers, archive each polygon as GeoJSON with effective dates. Autonomy regressions should pin map version like they pin model checkpoint.
What we did not verify this run
We did not independently geocode Tesla's map image or pull the X post ID for embed. Use Robotracker and EVwire as secondary checks; Tesla's account remains the primary boundary announcement when you need legal certainty.
Insurance and municipal fee watchlist
Austin, Travis County, and Texas state motor-vehicle rules can change permit fees independently of geofence posts. If you operate a fleet analytics startup, model fee sensitivity when Tesla or Waymo polygons overlap suburbs that impose per-trip taxes. Our DC Waymo cap fight shows how VMT fees follow visible robotaxi growth.
Riders should screenshot trip receipts with pickup pins when debugging out-of-zone billing disputes — geofence bugs show up as edge-case polygons, not global outages.
Waymo still runs a smaller Austin footprint on Robotracker; competitive riders in the same metro may see different apps, prices, and vehicle types on the same street corner — compare experiences, not press release acreage alone. Volvo and NVIDIA's HALOS narrative is a third lane — consumer highway stacks — that does not compete on city geofence area at all.
Journalists should always print mi² and km² beside any 443 figure copied from social posts — the number is often square kilometers from 2025, not square miles in November 2026. Robotracker's 435 mi² reading on November 9 is the better anchor for today's story. EVwire's 440 mi² estimate agrees within normal map-scaling error for metro-wide polygons.