TSL Robotaxi is on the road, by invitation only, 30 yuan per ride, even if you have money, you can't experience it.

Author: siqi

After 10 years of Elon Musk talking about autonomous driving, Tesla's Robotaxi is finally on the road.

On June 22 local time, Tesla's Robotaxi officially launched services in Austin, the capital of Texas, USA. The initial deployment of Robotaxis is approximately 10-20 vehicles, and in the early stages, a select group of investors and social media KOLs will be invited to participate in the experience.

As the dual-door "True Autonomous Vehicle" Cybercab released last year has not yet officially entered mass production, the first batch of Robotaxi services will be provided by the currently produced Model Y. This means that the current Robotaxi is still equipped with a steering wheel and accelerator and brake pedals, and each vehicle also has a safety officer in the passenger seat, with some geographical restrictions on the operational area.

Musk stated on the X platform that the fixed cost for the first batch of passengers is $4.20 per trip, but this is more like a "trial price," with long-term costs unknown. Currently, in Austin, Google's autonomous driving company Waymo has also launched Robotaxi in cooperation with Uber. According to previous reports from TechCrunch, the cost for Waymo is $3.50 per kilometer in the range of 4.3 to 9.3 kilometers.

Tech blogger shares Robotaxi travel experience: 16 kilometers, 22 minutes | Image source: X

Compared to the mainstream Robotaxi on the market, Tesla's problem-solving approach has two core differences:

First, compared to the current mainstream Robotaxi solutions such as Waymo and RoboTaxi, Tesla is the first to use "mass-produced vehicles" + FSD software for autonomous vehicle operations without the need to install complex perception hardware or redevelop new software. This is expected to significantly reduce both hardware and R&D costs.

Second, Tesla is attempting to prioritize attracting more "car owners" (rather than "users") through the operation of the first batch of Robotaxis. If this model can work, then Tesla will not need to spend its own money to buy cars and then put them on the market, but is expected to directly transform from a traditional state-owned taxi company into a ride-hailing platform like Didi and Uber.

Therefore, despite some "regrets" in form, it does not affect the fact that the Tesla Robotaxi remains the most关注 event in the entire autonomous driving and driverless vehicle industry.

Currently, Tesla's market value is $1.01 trillion, with a price-to-earnings ratio of 177, far exceeding traditional automakers. In recent years, Musk has repeatedly emphasized at various occasions that Tesla "has transformed from a car company into an AI company," and the launch of Robotaxi in the market is an important demonstration of its capabilities and products. A model from the famous investment firm Morgan Stanley shows that in Tesla's future valuation, 60-80% comes from expectations regarding Robotaxi and FSD business.

"driverless car" with a steering wheel and a safety officer

Most people’s expectations for Tesla's Robotaxi are based on the "driverless car" Cybercab that Musk unveiled last October at the Warner Studios in Hollywood. That vehicle completely eliminated the design of the steering and acceleration/braking pedals, featuring a two-door, two-seat design, with a screen in the front row serving as the core method of human-machine interaction.

Therefore, most people may fall into Musk's verbal trap, subconsciously linking Cybercab and Robotaxi. But in fact, Cybercab hasn't even started mass production. The Robotaxi launched in Austin this time is serviced by the 2025 Tesla Model Y. This vehicle is equipped with Tesla's HW4.0 (also known as AI4) smart driving hardware and software solution, which adopts pure visual perception, with a total of 8 cameras and Tesla's self-developed chip, providing a 360° surround view.

From the currently released test ride experience video, it can be seen that each Robotaxi is equipped with a safety officer in the front passenger seat. According to reports from American media, Tesla has also established an electronic fence for the Robotaxi, allowing only invited passengers to book the vehicle, and currently, it is only available in a small portion of the southern area of Austin.

Tesla Robotaxi buyer showcase (top) and seller showcase (bottom) | Image source: X

Why do such changes occur between the previous and current models?

Firstly, from a technical perspective, Cybercab still needs to overcome a series of challenges before large-scale production.

According to the timeline revealed by Musk at the earnings communication meeting, it is expected to officially launch before the end of 2026. Additionally, according to current regulations, vehicles without steering wheels and pedals require separate exemption approval before going to market, and a maximum of 2,500 units can be manufactured each year.

Secondly, from a marketing perspective, Cybercab primarily serves as a display function.

Just like Xiaomi's prototype car and the production version of the SU7 Ultra brushed the fastest lap in New York, respectively. Xiaomi doesn't need mass consumers to understand the subtle differences between prototypes and production cars, nor does it want to sell Xiaomi's cars only to users who are willing to go on the track, but to emphasize the capabilities of the brand and products through this matter: Xiaomi's electric cars have a sense of driving and are faster than Porsche.

The significance of Cybercab to Tesla is that it proves the FSD software can drive itself anywhere. Moreover, the absence of a steering wheel and pedals actually represents the metaphor that "anyone in the world can drive."

Traditional car manufacturers tend to build brand perception through speed, but supercars are often not the most mass-produced models. Cybercab is the "supercar" of Tesla as a smart car brand in this era—despite costing only $25,000, it supports over 60% of Tesla's extremely high market value.

So, if one day in the future, Musk tells everyone that Cybercab will not be manufactured in large quantities, please do not be surprised, because in Musk's plan, its role may not be about volume at all.

There is a prominent Robotaxi lettering on the side door of the Model Y | Image source: X

Robotaxi: A business that needs to find new problem-solving approaches.

This time, Tesla only deployed 10-20 Robotaxis in a small area of Austin. It may not seem large in scale, but why is it still making headlines in the tech world?

One major reason is that this may provide a new solution approach for the currently money-burning Robotaxi business.

In 2024, Waymo, the largest Robotaxi player in the U.S. market, provided over 4 million services, with a fleet size of about 1,500 vehicles, generating revenue of 50-70 million dollars, but incurred a terrifying loss of 1.5 billion dollars.

The revenue and investment are fundamentally disproportionate, which is the biggest pain point for the Robotaxi business at present.

Revenue is constrained by the current fleet size and market demand, with investments mainly divided into two parts: software, primarily focused on the development of autonomous driving features, and hardware, which includes vehicle purchases and hardware installations.

Unlike Tesla, the mainstream Robotaxi solutions on the market today achieve functional and safety redundancy by adding perception hardware to mass-produced vehicles. This set of redundant hardware is often composed of lidar and high-performance autonomous driving chips, and the BOM cost of just these hardware components can reach tens of thousands or even exceed one hundred thousand RMB.

Waymo's active vehicle models are equipped with an autonomous driving safety system on top of the vehicle | Image source: Waymo

Secondly, the current Robotaxi companies are actually more like traditional bus companies. They need to spend money to buy cars first, then hire "people" to drive (which includes software development, as well as costs for safety personnel and operations), and only then can they push to the market.

It is evident that this process is somewhat redundant. Using the "traditional model" for Robotaxi (which is not traditional, as it is a new technology) requires too many investments: it is necessary to invest in R&D like an AI company, buy equipment like a hardware company, and possibly provide low-cost subsidies to consumers during operations.

Elon Musk's ideal self-driving car business does not rely on building taxis himself, but rather on using owners' cars to make money for Tesla.

Musk once said that in the future, he hopes Tesla's fleet of autonomous taxis will reach a global scale of 1 million.

At a cost of $25,000 per vehicle, the hardware manufacturing cost for these cars alone reaches $25 billion, more than three times Tesla's net profit for the entire year of 2024. Clearly, if Cybercab is mass-produced before entering the market, even with regulatory approvals in various global markets being smooth sailing, this project would still represent a massive investment.

So, in fact, the 10 Robotaxis this time can essentially be understood as "test drive cars." This is yet another round of marketing activity: last year's Cybercab was a demo presentation, and this time the 10 cars are meant to prove the mass production capability of Tesla's FSD software.

American netizens share Tesla Robotaxi, pure vision technology route to cope with nighttime environments | Image source: X

Elon Musk's true goal is to prove that the mass-produced Model Y and Model 3 have the capability to directly become Robotaxis. If mass-produced vehicle owners are willing to join Tesla's Robotaxi fleet, then Tesla can transform from being a "traditional bus/taxi company" that sells cars into a "Tesla-branded Didi/Uber" mobility platform company in the autonomous vehicle business. Moreover, it can collect money twice: once when selling the car and again by taking a cut when accepting rides.

Pouring cold water: Technology Ready is not enough

In the valuation system of the renowned investment firm Morgan Stanley, over 60% of Tesla's future value comes from FSD and Robotaxi.

At ARK, where the famous investor "Cathie Wood" is located, the future value contribution of Robotaxi even exceeds 90% | Image source: ARK Invest

As mentioned earlier, Tesla aims to directly connect the entire process from L2 to L4 through mass production technology, relying on two new business models besides selling cars: Robotaxi and FSD (software subscription).

Thanks to the current global sales and ownership of the Model Y and Model 3, theoretically, any Tesla owner with HW4.0 hardware or above can turn their car into a "Tesla-branded Didi car" during idle times. This way, the million-vehicle fleet scale that Musk has always emphasized doesn't seem to be a problem.

However, even if we assume that the technology, operations, and approval details all go smoothly, the question is:

Do Tesla owners really all do this?

For old car owners, are they willing to use their private vehicles to provide public services?

For new car owners, is there anyone willing to specifically buy a Tesla to earn money by driving it as a taxi?

Taking the Chinese market as an example, the price of a new Tesla Model Y equipped with HW4.0 starts at 263,500 yuan. According to the current travel market rules, this car can be投入 the "dedicated car" market. However, as a new product, Robotaxi will likely need to adopt pricing subsidies and other measures to compete for user mindshare based on experience from the ride-hailing market. So, who will bear this cost?

And then there's the classic question: Musk always says that people can let their Model Y / Model 3 go out to pick up jobs when they don't need to use the car, but at that time, maybe no one has the need for a ride (for example, at midnight).

Therefore, from the perspective of pouring cold water on things, Robotaxi and autonomous driving are still businesses that cannot be achieved overnight. However, from an investor's perspective, even though Tesla's pure vision mass production solution has not been fully validated, it is still worth taking a "gamble" on the future due to its extremely high cost advantage.

This is also one of the key reasons why Tesla's price-to-earnings ratio remains as high as 177 times this year, after experiencing a series of hiccups, which is three times higher than any of the seven giants in Silicon Valley.

Before Robotaxi achieves results, Tesla and Musk's primary task is still to "sell cars," selling cars that are not Robotaxi.