Can Tesla Cybercab Actually Kill Uber? Day-Two Pricing From Austin

Day two of Cybercab in Austin. Farzad starts at The Domain on the north tip of Tesla’s Robotaxi geofence, about twenty minutes from downtown, and books a ride to the University of Texas Department of Neuroscience, roughly eight miles away. The point of the tape is not another cabin tour. He wants a live price and wait check against Uber, then the longer argument: can a purpose-built car with no steering wheel and no pedals actually put Uber in a corner once the fleet is thick enough.

The Domain price check

In the Robotaxi app he can pick Model Y or Cybercab. Model Y shows about a seven-minute wait at $19.68. Cybercab shows about thirty minutes at $14.87. He books the Cybercab. On Uber for the same destination, UberX is about $25 with a two-minute wait, Comfort about $32 with a one-minute wait, and Premiere about $56. Uber wins on wait. Cybercab wins on price in this single snapshot. Farzad’s read is that the long Cybercab wait is launch congestion: creators are burning rides around the public window. Tesla Robotaxi first posted public Cybercab rides for Friday, Sept 4, 2026 at 5:00 p.m. CT, then moved that clock up to 2:00 p.m. CT the same day, citing popular demand. On camera he still uses the original 5 p.m. line. Tesla’s fix, in his words, is more Cybercabs on the map, not a permanent half-hour queue.

He also notes the map tell. Early Austin Robotaxi started as a small south-Austin pocket. Cybercab, he says, is already bookable across the full geofence, including this north-edge Domain pickup. Tesla’s own site at launch still described Cybercab rides as limited parts of Austin; his Domain-to-UT booking is the map tell on this tape. He puts the Cybercab count around forty units in the fence, versus something like ten to twelve Model Ys at the original Robotaxi open. Treat those as his estimates from the ride, not a Tesla filing. Texas DMV listings around launch week put 45 Cybercabs on Tesla’s statewide authorized roster. The June 2025 Model Y open was reported in the 10 to 20 range.

Why Uber’s economics look broken to him

The car has no driver controls. Farzad’s cost story is simple: you are not paying for a human’s time, car payment, fuel or energy markup, and maintenance the way an Uber fare has to. Long term, once the safety case is strong enough for anywhere-anytime driving, he expects cost per mile for a car like this under 40 cents, and closer to 20 cents if remote supervision stays light. Those are his targets, not a published Tesla P&L line.

Uber’s answer, he walks through, is to put other companies’ robotaxis on the Uber app. He says Tesla has already drawn the line: Cybercabs stay on Tesla’s Robotaxi network. You do not hail a Cybercab inside Uber. So Uber needs Waymo, Zoox, Geely, Baidu, or whoever else will supply driverless metal. Even if Uber matches Tesla’s unit dump over time, Farzad’s squeeze is price. A private driverless ride on Tesla’s app at the lower mile rate leaves Uber needing either a premium people will pay, or a price cut that pays human drivers less. Lower driver pay thins the human network that made Uber’s waits short. Unless AV supply arrives at the same speed drivers leave, the marketplace gets thinner in the gaps the machines do not cover yet.

He stacks one more cost against Uber. Tesla owns manufacturing, the driving stack, and the customer app. Uber, in his framing, pays for someone else’s car and still needs a markup for itself. Matching Tesla on price while paying both would mean losing money. He also claims Waymo’s cars are Chinese now and that American and European OEMs still cannot match that EV build. Mark that as Farzad’s claim in the video, not a settled supply-chain audit. Waymo’s next-gen Ojai is built on a Zeekr/Geely platform from China; the older Jaguar I-Pace fleet is still in service.

Politics, fleet partners, and the factory problem

Farzad is blunt about the job hit. Driverless miles destroy driving wages. He is sympathetic to that. His counter is that trying to freeze people in a job the machine will take anyway only buys time for owners, not workers. He frames Uber’s rational play as lobbying and union politics to slow AV rollout, and he argues with Jason Calacanis’s All-In line that job loss will feed violence against Waymos and Teslas the way data-center backlash already feeds AI anger. That fight is Farzad’s opinion and his read of Calacanis, not a policy forecast.

Then the rollout problem. If regulators or politics keep Cybercab out of cities while the factory wants to run, Tesla still has to place metal. Farzad’s valve idea, which he says Tesla community types mock him for, is a steering-wheel-and-pedals variant of the same chassis sold to the public so the line does not sit at 5 or 10 percent. Test units with wheels and pedals did exist; production Cybercabs on Austin streets do not. Separately, Tesla posted a public interest form at tesla.com/robotaxi/interest on Sept 3, 2026, with checkboxes for Cybercab fleet purchasing, mobility hubs and infrastructure, event collaboration, and other. Farzad reads that as Tesla not wanting to own cleaning, repairs, and fleet ops at million-unit scale. Sell the car to a fleet operator, keep those cars on the Robotaxi network, take a network cut. That network-stay rule is his read of Tesla’s own-app model, not a line on the form. He floats a hypothetical 30/70 Tesla-to-operator split as an example, not a published rate.

What the ride itself showed

The Cybercab arrives with the app’s green glow and butterfly doors. He closes the door from the phone, starts the trip, and talks through urban and campus traffic on the way into UT. No highway yet on this product, he says. He wants that next. On exit, the car shows rear and side camera views so he can leave without walking into traffic. Empty car, no wheel, no pedals. It waits a beat after he gets out, then handles a crosswalk politely while he films from the curb.

One Domain-to-UT quote is not a national Uber funeral. It is an early public-window price and wait sample inside one geofence, plus Farzad’s stack-ownership thesis. If Tesla can keep dumping purpose-built cars onto its own app at a fare human networks cannot live on, Uber’s problem is not a better tip screen. It is owning neither the factory nor the customer once the wait times catch up.

This Exclusive is from the long-form at https://www.youtube.com/watch?v=IsvlE6ubxcs.

---

Receipt

Event → Farzad proof → book beat → buy link.

• Event: Day-two Cybercab public-window pricing from The Domain to UT Austin Neuroscience.
• Farzad proof: Live app quote Cybercab $14.87 / ~30 min vs Model Y $19.68 / ~7 min vs UberX ~$25 / ~2 min; full-geofence book from the north edge; stack + fleet-partner thesis.
• Book: AoC Ch.2 — Paid miles without a human in the loop is the product, not a demo. Master Plan Autonomy — who owns the network when the wage leaves the seat.
• Get the books: Abundance or Collapse · Master Plan