First Week of Tesla’s Cybercab Passenger Service: What Do Early Riders Think?

Elon Musk with a Tesla Cybercab

Tesla CEO Elon Musk with the Cybercab. The image is AI-processed.

By Li Hailun | Edited by Xu Qingyang

Tesla’s Cybercab has begun making trips with passengers on board.

On September 4 local time, the first Cybercabs entered Austin’s Robotaxi fleet and started offering paid rides within a limited local area. For now, riders hail a car through the Robotaxi app and cannot choose a Cybercab; vehicles are assigned by the system.

The arrival of the Cybercab in the fleet comes as Tesla’s Robotaxi network continues to expand.

Over the past year, Tesla has gradually widened service in Texas and Florida. This year it added Miami, Orlando and Tampa, and several cities have run unsupervised from day one. According to the latest registration records, Tesla now has 420 autonomous vehicles registered in Texas, including 375 Model Ys and 45 Cybercabs.

Putting the Cybercab on the road is only the first step. What matters next is how the passenger experience looks, how large Tesla’s Robotaxi network has become, and whether the model can grow beyond a few dozen vehicles into scaled operations.

01 — Cybercab begins to face passenger scrutiny

The Cybercab’s interior is extremely simple: two seats, one large screen, and no steering wheel, accelerator or brake pedal. Butterfly doors open automatically. Once a passenger sits down, the vehicle adjusts the seat and climate control according to personal settings, and it can even continue playing the music or video that was playing on the passenger’s phone.

Tesla Cybercab interior

Passengers who have already ridden in a Cybercab mostly give feedback in two areas: the car drives smoothly, and the interior and seating experience are quite different from an ordinary car. Meanwhile, several details involving hailing, locating the car and in-cabin controls still need adjustment.

Tesla owner and content creator Sawyer Merritt has spent about five hours riding in Cybercabs, with his longest trip roughly 40 minutes. He says there was no obvious jarring, hesitation or odd behavior during the rides, and after several trips he finds the car drives very steadily overall.

Sawyer Merritt riding in a Tesla Cybercab

His suggestions focus mainly on cabin features: a full-screen map, the ability to change the destination directly on the screen, a rear camera view, and manual control of air recirculation and the windshield washer.

Sawyer Merritt Cybercab ride video

Merritt’s ride video. Source: X

Devin Olsen has taken more than 30 rides. The issues he raised: when several Cybercabs are parked together, he wants the cars to be easier to identify through lights or similar cues; he wants support for multiple drop-off points; and he hopes the map offers a satellite view and that riders can adjust the drop-off point directly inside the car.

Devin Olsen Cybercab ride video

The ride video Olsen posted. Source: X

These responses have a common thread. Getting passengers from pickup to destination is already the baseline capability. Once the car moves into everyday use, passengers start paying attention to the details a human driver once handled: finding the car, changing the destination, and getting guidance after getting out.

That point is even clearer in the experience of Gary Moss, a visually impaired passenger. Moss lost his eyesight five years ago. Before that, he drove about 31,000 to 40,000 miles (50,000–65,000 km) a year. After losing the ability to drive, even going to a coffee shop, attending activities, or handling daily errands often required family accompaniment.

During his Cybercab ride, the automatic doors and the spacious cabin made a deep impression on him. To him, the cabin felt like “his own living room” where he could sit, listen to music, and chat with his son while the car finished the remaining journey.

Gary Moss in a Tesla Cybercab

But for Moss, finding the car and getting guidance after getting out matters even more. For instance, when several Cybercabs are parked together, there is no obvious sound cue that helps him find which car is his. Grok can answer questions about the current location, remaining distance and arrival time, but it still cannot fully replace a phone for setting the destination.

Moss wants the car to announce, on arrival, where it has stopped and how much farther passengers need to walk from the stopping point to their destination.

As the Robotaxi opens up to ordinary passengers, the experience standard is also changing. In the past, assessing an autonomous vehicle centered on whether it could finish the drive. Now passengers ask more: Is it easy to hail the car? Is the car easy to find? Can the destination be changed at any time? And after getting out, will passengers still get enough help?

The Cybercab is now beginning to be tested against those questions.

02 — The economics of the Cybercab

Since the design stage, the Cybercab has made trade-offs around Robotaxi operating costs.

The car uses a 48 kWh battery pack with 4680 cells and has a drag coefficient below 0.20. The electric drive system is smaller and lighter. On the manufacturing side, Tesla adopted an “unboxed” production approach, hoping to shorten the assembly line by roughly 50%. The body also uses reaction-injection-molded plastic, further reducing traditional metal-body manufacturing and paint work.

Musk previously set targets for the Cybercab: a price below $30,000 and an operating cost of around 20 cents per mile. After taxes and other expenses, total cost would be roughly 30 to 40 cents per mile.

For a Robotaxi, however, whether costs can really fall also depends on another metric: vehicle utilization.

One direct measure of utilization is how many trips a car can handle in sequence during a day. Luis Ovalle, a user in Spain, gave an example of a future Cybercab routine: in the morning, the Cybercab drops him at the office and then drives itself home; a little later, it takes his wife to the airport, then picks up his daughter from the university, and finally returns to the office to pick him up.

Luis Ovalle’s envisioned Cybercab usage scenario

Ovalle’s envisioned Cybercab usage scenario. Source: X

In his view, a single car can take on different members of a family’s travel tasks one after another, while people inside the car can work or do other things.

This usage is also part of the fleet model Musk has described. Under that model, owners use the car themselves; if they go away for several days, they can add it to the fleet through the app and recall it when they need it. Tesla will operate some vehicles itself and will also allow end users to put their own vehicles into the fleet.

Such a model requires high vehicle utilization. A normal private car sits parked for large parts of the day, but once it joins the Robotaxi network, it can keep accepting trips while the owner is not using it. The more vehicles there are, the higher the requirements for dispatch, maintenance, charging and operating systems.

03 — Live in seven cities, still on the eve of the big test

Before the Cybercab arrived in Austin, Tesla’s Robotaxi service already covered seven markets: Austin, Dallas, Houston, Miami, Orlando, Tampa and the San Francisco Bay Area. For now, the Cybercab offers rides only in Austin, where it is mixed into the existing Robotaxi fleet.

Map of markets where Tesla Robotaxi operates

The rollout pace differs from market to market.

Austin, Dallas and Houston entered Robotaxi operations earliest. Miami, Orlando and Tampa came online in July this year, and Tesla adopted unsupervised operation directly in those cities.

The San Francisco Bay Area still requires a safety driver. The service runs under California’s TCP commercial passenger permit, making it not a fully driverless Robotaxi deployment.

At the second-quarter earnings call on July 22, Tesla’s AI software vice president, Ashok Elluswamy, said cumulative unsupervised mileage in two states across six cities, excluding the Bay Area, had exceeded 373,000 miles (600,000 km). Tesla also said that in the second quarter it continued expanding its unsupervised operating area in Austin.

The expansion order shows that Tesla first chose markets with relatively relaxed regulatory conditions. Texas places fewer restrictions on autonomous commercial operations, while Florida’s state-level policy leaves substantial room for autonomous ride-hailing. In those places, Tesla can limit its operating area with geofences and then expand gradually.

California is different. Tesla still needs safety drivers in the Bay Area, and its autonomous-driving business is subject to several layers of oversight from the DMV, CPUC and other agencies.

Tesla Robotaxi rollout and regulation

Next up are Phoenix and Las Vegas. Tesla has already listed Phoenix as “preparing.” Las Vegas received approval from Nevada on August 20 to deploy up to 5,000 fully autonomous vehicles in its first year. But before officially carrying passengers, the company must complete vehicle inspections, insurance, fare filings and app review. Airport service also needs separate approval.

The market list looks long, yet measured against the entire U.S. mobility market, Tesla’s Robotaxi remains at an early stage. The company has not disclosed its full fleet size or weekly order volume. Outside estimates also use different definitions: some count registered vehicles, some count vehicles in active operation, and some count only driverless vehicles.

What can be confirmed right now is that the Cybercab still accounts for a relatively low share of Texas’s autonomous fleet. Its arrival shows that the purpose-built Robotaxi has entered commercial operation, but it is still far from forming large-scale capacity.

The immediate task for Tesla is to scale an operating model that has already been validated across far more vehicles.

04 — Hurdles the Cybercab still must clear

The number of Cybercabs actually in commercial service remains very small, still far from the more than 125,000 vehicles of annual capacity Tesla has installed. Manufacturing, software, regulation and operations all lie in between.

In early September, the National Highway Traffic Safety Administration (NHTSA) opened a review of Tesla’s self-certification process for the Cybercab, involving an estimated maximum of about 1,000 vehicles. The review centers on the technical materials and certification rationale Tesla used to demonstrate compliance with federal motor vehicle safety standards, including whether the company concluded that some traditional safety standards do not apply to the Cybercab.

Passenger rules also show the practical boundaries of the current Robotaxi business. Under Tesla’s current rules, customers aged 18 and above can book and ride alone. The Model Y Robotaxi allows minors aged 8 to 17 to ride only if an adult accompanies them for the entire trip; children under 8 are not allowed.

The Cybercab has a higher age cutoff: minors aged 13 to 17 may ride with an adult, while children under 13 are not allowed. For child safety seats and booster seats, passengers must follow local requirements and prepare and install the equipment themselves.

Cybercab passenger rules

Tesla has not set up a separate Robotaxi policy for students, nor has it announced campus commuter service or student discounts.

For the Cybercab, 45 vehicles are only the first batch in the operating network. What matters in the future is how quickly produced cars can enter the fleet, how many orders each car can complete per day, and whether driverless operation can keep spreading from a few bounded areas to more cities.

If this stage can be worked out, automakers’ revenue structures will also change. After a car is sold, it can stay in the operating network and keep serving passenger trips.

For Tesla, the Cybercab story has just moved from “building it” to “running it.”

Related Articles

分享網址
AINews·AI 新聞聚合平台
© 2026 AINews. All rights reserved.