With Robotaxis on the Road, What's Behind Tesla's Bet Against Lidar?
Sign up for ARPU: Stay ahead of the curve on tech business trends.
Tesla rolled out its long-promised robotaxi service in Austin this week, a landmark moment CEO Elon Musk called the "culmination of a decade of hard work." The deployment of driverless Model Ys carrying paying customers puts Tesla squarely in the autonomous ride-hailing market. But within hours of the carefully managed launch, videos of the cars driving erratically, braking hard, and even entering a wrong-way lane went viral, immediately drawing the attention of the National Highway Traffic Safety Administration (NHTSA).
The rocky debut throws a spotlight on the fundamental schism that divides the multi-billion-dollar autonomous vehicle industry. It’s a high-stakes debate between two competing philosophies: Tesla’s bet on cameras and artificial intelligence alone versus the rest of the industry’s reliance on a full suite of expensive sensors. With trillions of dollars in potential market value on the line, the performance of these first robotaxis in Austin is a critical test of whether Musk’s audacious gamble will pay off.
What happened at the Austin launch?
Tesla’s robotaxi service began on June 22 in a limited, geofenced area of Austin. The initial trial involves a small fleet of about 10 Model Y vehicles, available by invitation only, primarily to social-media influencers and Tesla supporters. The cars operate without a person in the driver's seat, but do have a "safety monitor" in the front passenger seat. Musk stated on his social media platform X that the company was being "super paranoid about safety," with rides offered for a flat fee.
Despite the controlled conditions, several videos posted by the public showed the robotaxis exhibiting confusing behavior. One was seen driving briefly on the wrong side of a road, while another braked sharply in traffic for stationary police cars outside its path. On June 23, NHTSA confirmed it was "in contact with the manufacturer to gather additional information" about the incidents.
How does Tesla's approach differ from its rivals?
Tesla is bucking industry convention by pursuing a "vision-only" strategy. Its vehicles rely on a set of eight cameras, processed by a powerful onboard computer and a sophisticated AI neural network, to "see" and interpret the world, much like a human driver. Musk argues this approach is not only more elegant but also more scalable, as the hardware is relatively cheap and already exists in millions of Teslas on the road, which are constantly gathering data to train the AI.
Nearly every other major player—including Alphabet's Waymo and Amazon's Zoox—uses a "belt-and-suspenders" approach. Their vehicles are equipped with a full sensor suite that includes cameras, radar, and, most critically, lidar. Lidar (Light Detection and Ranging) uses lasers to create a precise 3D map of the car's surroundings, providing a layer of redundancy that works in conditions where cameras might struggle, like darkness or direct glare. As Waymo's co-CEO Tekedra Mawakana has argued, the goal is to "replace humans," which requires a system with vision that's "better than humans," not just mimicking them.
What are the arguments for each strategy?
The core of Tesla’s argument is cost and scalability. Without expensive lidar and radar units, which can add tens of thousands of dollars to a vehicle's cost, Tesla can theoretically turn any car it produces into a robotaxi with a software update. Analysts have estimated a Tesla robotaxi vehicle costs a fraction of a Waymo equivalent. This is the foundation of Musk's vision for a fleet of over a million robotaxis, consisting of both company-owned cars and personally-owned Teslas added to a ride-hailing network.
The argument for the multi-sensor approach used by Waymo is safety through redundancy. In a 2025 peer-reviewed study of over 56 million miles of its operations, Waymo reported 85% fewer crashes with suspected serious or worse injuries compared to human drivers in the same areas. Proponents argue that relying on multiple, different types of sensors is the only responsible way to handle the infinite "edge cases" of real-world driving. This stands in contrast to Tesla's Autopilot and Full Self-Driving systems, which have been the subject of dozens of NHTSA investigations following crashes.
So, who is actually winning the robotaxi race?
By commercial metrics, Waymo is the clear leader. As of mid-2025, the company was providing over 250,000 paid rides per week across four cities and had surpassed 10 million total paid trips. It has built a slow, steady, and largely safe operational history.
Tesla, however, is the undisputed leader in potential scale. While its robotaxi service is just beginning, it has already sold millions of FSD-capable vehicles. If it can perfect its "unsupervised" software, it could activate a massive fleet almost overnight. This potential is a key reason for Tesla's sky-high valuation. Wedbush analyst Dan Ives has argued that the robotaxi business alone could add a trillion dollars to the company's value.
Meanwhile, a robust ecosystem is thriving in China, where companies like Baidu and Pony.ai are operating in multiple cities with strong government support, representing a formidable challenge to U.S. dominance.
What are the biggest hurdles ahead?
The Austin launch highlights the primary challenge for all players: closing the gap between a system that works 99% of the time and one that is safe enough for all conditions. The final percentage points of reliability are proving to be the most difficult and expensive to achieve.
Regulation remains a major wildcard. The new Texas law signed just before Tesla's launch signals that even business-friendly states are moving toward more oversight. A national rollout in the U.S. will require navigating a complex patchwork of state laws, with places like California demanding far more stringent testing and data reporting.
Finally, there is the court of public opinion. High-profile incidents, whether from GM's Cruise in 2023 or the viral videos from Austin in 2025, can damage public trust across the entire industry. As protests in Austin ahead of the launch demonstrated, convincing people to trust a robot with their lives is a challenge that may be even harder to solve than the technology itself.