Image: Tesla/ X
Tesla has cautiously initiated its long-anticipated robotaxi service in Austin, Texas, launching on Sunday with approximately 10 Model Y SUVs under strict operational guidelines.
<pAccording to CEO Elon Musk's post on X, rides were offered at a fixed price of $4.20, and social media influencers were observed booking and experiencing rides in various locations throughout Austin, as shown in videos shared by Musk.
Musk noted that the company is adopting an extremely cautious approach to safety, with human operators remotely monitoring the fleet, alongside safety features in the front passenger seats of the vehicles.
This remote control capability, referred to as “teleoperation” in the industry, is employed to varying extents by several robotaxi startups across the world. While it presents distinct benefits, it also has significant limitations.
Congratulations to the @Tesla_AI software & chip development teams on a successful @Robotaxi launch!!
A testament to ten years of diligence.
Both the AI chip and software teams have been formed from scratch within Tesla.
— Elon Musk (@elonmusk) June 22, 2025
Here’s an overview of how it functions:
Understanding Teleoperation
Teleoperation refers to the control of machines from a different location, typically via a wireless network.
This technology is utilized for training robots to function independently, monitoring their autonomous actions, and intervening when necessary.
Application of Teleoperation by Robotaxi Operators
The global robotaxi sector is still in the testing phase, with companies deploying vehicles in limited areas while continuously fine-tuning the AI software that powers them.
Teleoperation frequently comes into play when a vehicle is uncertain about its next move.
For instance, Alphabet’s Waymo has a team of human “fleet response” agents who assist the Waymo Driver—its autonomous technology.
As explained by Waymo in a blog entry last year, “Similar to a friend on the phone, when the Waymo vehicle encounters a specific situation, the autonomous driver can contact a human agent for further clarification.”
Former Waymo CEO John Krafcik mentioned to Reuters, “The vehicles aren’t actively monitored,” indicating that the software ultimately makes the decisions.
A video from Waymo illustrates a situation where a car queries a remote operator about whether the road is clear of emergency vehicles. After receiving confirmation, it proceeds forward. Conversely, some companies like Baidu’s Apollo Go in China employ entirely remote backup drivers who can take control of the vehicles if needed. Baidu opted not to comment on this.
Challenges of Teleoperation
One of the principal challenges of driving remotely on public roads is dependency on cellular data connections, which can falter or exhibit delays, resulting in the loss of communication between the vehicle and the remote operator at critical times.
Philip Koopman, an engineering professor at Carnegie Mellon University and an expert on autonomous vehicle safety, stated that this method may work for a small-scale pilot project like Tesla’s in Austin but labeled teleoperation as “inherently unreliable.”
“Eventually, a disconnection will occur at the most inconvenient moment,” he noted. “While this might not happen with a fleet of 10 cars, it will likely daily with a million.”
Krafcik concurred, adding that the delay in cellular signals renders remote operation “very risky.”
Conversely, relying heavily on vehicles to request assistance and allowing the vehicle to be the primary decision-maker also holds risks since it does not ensure correct decision-making.
Waymo did not provide comments on its operational limitations.
Koopman pointed out that there are also restrictions on how many vehicles a single operator can effectively supervise.
A group of Democratic Texas lawmakers recently requested Tesla to postpone the launch of its robotaxi service until September, coinciding with the enactment of a new autonomous driving law. The lawmakers argued that delaying the launch is crucial for public safety and trust in Tesla’s operations.
Tesla’s Robotaxi Strategy
Musk has long promised that Tesla’s Full Self-Driving (Supervised) advanced driver assistance technology would eventually evolve into fully autonomous robotaxi operations. Earlier this year, he announced the introduction of a paid service in Austin backed by an “unsupervised” version of the software.
“We will have Teslas on the roads, operating autonomously in June, in Austin,” Musk assured analysts and investors in January. In May, he informed CNBC that the robotaxi would be limited to specific parts of intersections and rely on human oversight for monitoring.
The role of teleoperators in this context remains somewhat ambiguous, as insiders indicated that teleoperators would assist in scenarios where a robotaxi faces challenges, such as navigating a congested pedestrian zone.
Tesla has been hiring for teleoperation roles, emphasizing the need for individuals capable of remote control and oversight of autonomous vehicles and humanoid robots. The job postings describe responsibilities involving “performing complex and nuanced tasks remotely.”
Tesla has not responded to inquiries for further information.
“We are extremely vigilant regarding safety,” Musk stated in a post prior to the launch.