Robotaxis are autonomous vehicles, or AVs, that provide ride-hailing services without the need of a human driver — and they’re quickly catching on as technology improves. Typically electric, these driverless cabs have the potential to ease traffic congestion, lower emissions and save money for passengers and companies alike, disrupting traditional taxi services and even the demand for personal cars in the long run.
Robotaxi Definition
Robotaxis are self-driving vehicles that provide on-demand transportation. Like traditional taxis, they can be hailed at-will through an app and shuttle passengers to their destination, but they do so autonomously — with no human driver behind the wheel.
With pilot programs already underway by industry leaders like Waymo, Baid and Zoox, self-driving taxis are expected to become a major player in future urban mobility, providing a more convenient and affordable way for people to get around.
What Is a Robotaxi?
A robotaxi is an autonomous, self-driving vehicle designed to pick up and drop off passengers on-demand. They are outfitted with a multitude of sensors, cameras, radar, LiDAR technology and audio receivers in order to navigate roadways, including traffic, obstacles and unpredictable behaviors known to urban streets.
Robotaxis are often associated with electric vehicle technology, as many of the models are either retrofitted EVs, hybrids or boxy new designs specifically designed for this purpose. They work similarly to ride-hailing services, like Uber and Lyft, but without the involvement of a human driver in the vehicle. Instead, passengers hail the vehicle using an app, and the robotaxi transports them to their destination — all while the onboard artificial intelligence controls the driving functions.
Robotaxi Examples
Waymo
Waymo, Google’s self-driving outfit under Alphabet, is a leader in the robotaxi space, and became the first-ever company to offer a driverless ride-hailing service to the public in 2020. To date, the company’s all-electric fleet of modified Jaguar I-PACE electric SUV and Chrysler Pacifica Hybrid minivans have collectively driven upwards of 100 million miles without a driver, offering more than 10 million paid rides. Waymo operates a 24/7 service in Phoenix, San Francisco, and Los Angeles, with plans to expand to Austin, Atlanta, Dallas and Nashville through a partnership with Lyft and Uber.
Tesla
First announced in 2019, Tesla has been pursuing a robotaxi service for several years, slowly making gradual but significant progress. In 2024, the company unveiled the Cybercab — a purpose-built, fully autonomous vehicle that doesn’t even have a steering wheel or pedals — but it isn’t expected to enter production until 2026. For now, Tesla’s very limited ridehailing services in Austin and San Francisco use Model 3s and Model Ys equipped with its Full Self-Driving software. However, because this software has not achieved Level 4 autonomy under regulatory definitions, a human must be in the driver’s seat for all rides to ensure safety.
Apollo Go
Multinational AI company Baidu operates Level 4 self-driving vehicles in over 16 cities, including Beijing, Wuhan,Shenzhen and Abu Dhabi. Its autonomous ride-hailing service platform, Apollo Go, utilizes LiDAR, cameras, radar and AI-based perception systems to navigate urban environments without human intervention. As of January 2025, Baidu’s car fleet has completed more than 14 million rides, making it one of the largest autonomous ride-hailing services in China. The company plans to continue its expansion through partnerships with Uber and Lyft, which will enable it to enter the European market in 2026.
May Mobility
May Mobility is an autonomous vehicle company based in Ann Arbor, Michigan. With a fleet of hybrid-electric Toyota Sienna minivans and custom-built shuttles, it has provided more than 500,000 autonomous rides, offering first-mile and last-mile transportation. May Mobility recently entered the robotaxi sector by offering fully driverless rides in Peachtree Corners, Georgia. With partnerships with Uber and Lyft the company is eyeing expansions into Atlanta and Arlington, Texas.
AutoX
California-based startup AutoX develops robotaxis at Level 4 and Level 5 autonomy. Its self-driving fleet, deployed in cities like Shanghai and Shenzhen, operates without human safety drivers in some areas with the help of LiDAR, radar and AI-powered perception systems to navigate dynamic traffic conditions. In 2022 AutoX took part in a grocery delivery program in San Jose, and became the second company ever to receive a permit to roll out fully driverless passenger vehicles in California.
Pony.ai
Pony.ai’s robotaxis, built from modified Toyota and Hyundai vehicles, are being tested in pilot programs for public ride-hailing in cities like Fremont, Beijing and Guangzhou within geofenced areas. They come equipped with complex sensor systems — LiDAR, radar and cameras — as well as machine and deep learning software to accurately predict on-road behavior. In 2022, Pony.ai became China’s first autonomous vehicle company to receive a taxi license and is on track to have a 1,000-vehicle fleet.
Zoox
Zoox, an Amazon subsidiary, is developing a fully autonomous, bidirectional robotaxi without a steering wheel, pedals or driver’s seat. These boxy, all-electric vehicles get around by way of cameras, LiDAR and radar, and can “see” over 150 meters in all directions — including around corners. With its self-driving system built into retrofitted Toyota Highlanders, Zoox has provided mobility-as-a-service in the San Francisco Bay Area since 2017, and have rolled out similar programs in Las Vegas and Seattle. Zoox also became the first domestic company to receive an exemption from the National Highway Traffic Safety Administration, allowing it to legally operate purpose-built autonomous vehicles on public roads.
EasyMile
French company EasyMile is making autonomous shuttles for first-mile and last-mile transport within controlled zones like business parks, campuses and airports. Carrying 10 to 15 passengers at a time, the company’s EZ10 shuttle operates at Level 4 autonomy, meaning it runs fully driverless on pre-mapped routes at low speeds of up to 25 miles per hour. Unlike traditional robotaxis, EasyMile focuses on shared, fixed-route transport rather than on-demand ride-hailing in urban environments.
How Do Robotaxis Work?
Robotaxis utilize machine learning and artificial intelligence to navigate dynamic, unpredictable roadways. Generally, these vehicles are equipped with LiDAR, radar, 360-degree cameras and ultrasonic sensors that detect obstacles, map surroundings and track nearby objects, creating a unified picture so it can “see” its way around. This is known as “sensor fusion,” a process that enables autonomous vehicles to make real-time decisions in a matter of milliseconds, Steven Spieczny, vice president of marketing at AI-software company Kognic, explained.
“If you have a camera on top of the car taking a flat 2D image, while a radar scans ahead of the car’s position to determine a 3D perspective and an infrared LiDAR sensor on the side of the car, scanning for thermal activity — all these feeds need to be correlated and calibrated,” Spieczny told Built In. “These sensors record everything going on within proximity to the vehicle,” he continued. They’re asking “What are these ‘objects?’ What are they doing — are they stopped? Moving? How fast are they moving? What is the risk?”
The robotaxi’s built-in AI system processes this data, using machine learning algorithms to identify objects like pedestrians, other cars, traffic signals and obstacles in order to adjust the vehicle’s speed, direction and route accordingly — and all while reacting to traffic conditions, road hazards and passenger requests. Continuously learning, the system constantly collects information and updates its understanding of its environment in real time, enabling the vehicle to operate with minimal human input depending on its level of autonomy.
Levels of Autonomy in Robotaxis
Starting at zero, there are six different measures of autonomous driving standards. Reliance on a human driver reduces as you move up the ladder, with Level 5 representing a completely autonomous system. Here’s a quick rundown of each stage:
- Level 0 — No Automation: The vehicle has zero automation capabilities, and the driver is fully responsible for all aspects of control. Most “regular” cars fit in this category.
- Level 1 — Driver Assistance: This level includes basic driver assistance systems, such as adaptive cruise control or lane-keeping assist. But the driver is expected to remain attentive and engaged, overseeing the vehicle’s operation at all times.
- Level 2 — Partial Automation: The vehicle can autonomously control steering, acceleration and deceleration simultaneously under certain conditions, such as with Tesla’s Autopilot. However, the driver must be ready to take control if necessary, and their hands should remain on the wheel.
- Level 3 — Conditional Automation: The system can handle all driving tasks in most conditions, such as highway driving, but the driver must be available to intervene if needed. The car can handle tasks like speed control, lane management and obstacle avoidance, and will alert the driver to take over when it encounters a situation it can’t handle.
- Level 4 — High Automation: All driving tasks can be performed autonomously within certain geofenced areas or conditions. For the most part, a human driver is not needed to intervene — unless the vehicle travels into an unmapped area or finds itself in an extreme situation, like severe weather or unexpected obstacles on the road.
- Level 5 — Full Automation: The vehicle is capable of full autonomy in all conditions, without the need for a driver at any point. Its capabilities exceed human intervention, so much so that these vehicles are often built without a steering wheel or pedals. These vehicles can drive anywhere under any conditions without human oversight.
Safety and Testing of Robotaxis
Robotaxis undergo thousands of hours of testing over the span of a few years to a decade before they can take to the streets. This process involves simulations, closed-track tests and real-world testing under controlled conditions, all of which is supervised by safety drivers to ensure the vehicles are safe and reliable enough for public release.
But these systems are far from perfect.
Cruise, a former frontrunner in the self-driving taxi industry, was banned following a series of incidents, including one where a pedestrian was dragged several feet by a Cruise vehicle. These events led to regulatory investigations, caused Cruise to halt services in other areas and prompted the company to shift its focus away from developing its robotaxi fleet, according to its parent company General Motors.
And Cruise isn’t alone. Self-driving cars have been involved in several fatal accidents, including a 2018 incident where an Uber vehicle struck and killed a woman, partly due to the car’s perception system misclassifying her as an object rather than a person. Computer vision systems in general can struggle to see in the dark or around corners. They’ve also been known to malfunction in unpredictable ways, misinterpreting stop signs as 45 miles-per-hour speed limit signs or failing to recognize an articulated “slinky” bus, leading to collisions.
Even so, human error still accounts for about 94 percent of all traffic accidents, according to the National Highway Traffic Safety Administration. And the most recent data finds that autonomous vehicles are “generally safer” than human drivers “under certain conditions,” Spieczny noted. For example, a Waymo spokesperson told Built In that collision data from more than 33 million passenger miles shows Waymo vehicles 81 percent fewer crashes with airbag deployment and 78 percent fewer injury-causing crashes than an average human driver.
The challenge is that one “rare event” involving a driverless car can significantly hinder public perception of the technology. The only way to solve this is to capture more data and train these vehicles to handle these unique cases in the future. Waymo, for instance, is refining its driverless technology for winter conditions, testing it in upstate New York, Michigan and Northern California to expand its capabilities, according to the company spokesperson.
Vivek Joshi, an industrial automation engineer who has worked on product launches for robotaxis and other autopilot-enabled vehicles, likens the process to a baby first learning to walk before it can run.
“You still fall no matter how good of a runner you are — it’s more about how frequent the malfunctions are,” Joshi told Built In. “Companies constantly train their autonomous systems to be better by the day. It’s just a matter of time when we … gain enough autonomy to a point where almost every vehicle will be following a similar system, resulting in the lowest probability of accidents.”
Benefits of Robotaxis
Self-driving taxis pose a number of perks, for both riders and companies alike.
Reduce Human Error
Human error is responsible for the vast majority of car accidents. Equipped with smart sensors and advanced algorithms, robotaxis have the potential to enhance road safety, ridding roads of drunk, distracted, rage-filled or fatigued drivers who are making poor decisions at the wheel.
So far, autonomous vehicles can react in milliseconds — a rate that either matches or beats the average human, ranging from about 0.4 seconds to 1.5 seconds. But whether driverless cars are actually “safer” is too early to call. On average, an article from the National Law Review estimated that self-driving cars are involved in about nine accidents per million miles driven, compared to four accidents per million miles for traditional vehicles.
Relieve Traffic Congestion
Carpooling robotaxis could eventually become a daily norm for riders, reducing the reliance on personal car ownership. This would put fewer cars on the road while allowing smart AVs to improve road efficiency — potentially up by 35 percent, a 2019 university of Cambridge study found — as they coordinate with one another, adjusting speeds to prevent stop-and-go patterns and avoid bottlenecks.
Also, consider that driverless taxis do not need to park near destinations, which would free up parking spaces and reduce blockages caused by parking — an inevitable nuisance that accounts for 30 percent of city congestion, according to the U.S. Department of Transportation.
Reduce Carbon Emissions
Robotaxis — especially electrically-powered models — offer many eco-friendly features, such as optimizing routes, minimizing idle time, promoting carpooling and the elimination of energy-sucking driving patterns, like excessive acceleration or braking.
Cut Costs
Driverless robotaxis can cut costs by eliminating the need for human drivers — one of the biggest expenses in ride-hailing. Research from McKinsey shows that autonomous vehicles could halve operational costs by 2030, meaning cheaper rides for passengers and higher profits for companies. And, by 2035, the industry as a whole could create $300 billion to $400 billion in revenue.
Challenges of Robotaxis
Even with impressive tech innovations and safeguards in place, robotaxis still need to navigate several off-road obstacles.
Safety Concerns
Autonomous vehicles are widely considered to be safer than having a human behind the wheel, but it’s a bit more complicated than that.
While Waymo and the now-defunct Cruise claim their vehicles are safer than human-driven cars, their reported crash rates raise questions. Waymo reported 4.5 crashes per million miles in 2023, compared to the national average of 1.94. Additionally, Cruise reported a higher crash injury rate of 1.0 per million miles at its Phoenix-based operations, which is above the average injury crash rate for drivers in that area. Keep in mind that these pilot programs are also skewed, as they are restricted to limited areas with favorable conditions.
Public Perception
People are still hesitant to jump in a self-driving car. In a 2025 survey by AAA, only 13 percent of U.S. drivers would trust to ride in a self-driving vehicle. Though it is nine percent higher than what was reported in 2024, it pales in comparison to India and China, where over 50 percent of drivers trust self-driving cars. While the U.S. may see early adopters, robotaxi companies may find it harder to reach mass adoption.
Regulatory Requirements
Laws that dictate what autonomous vehicles can and can’t do vary widely by region. In California, for example, companies must obtain the proper permits and licenses, report collisions in a certain way and ensure a safety driver is present in every test vehicle, regardless of its level of autonomy. In contrast, states like Arizona — incidentally, where the first robotaxi-related death occurred — are a bit more lax when it comes to regulations.
This lack of standardization will ultimately delay the widespread adoption of autonomous vehicles, stalling even the most successful AV companies as they scale.
The Future for Robotaxis
When it comes to the mainstream adoption of robotaxis, it’s no longer a question of if, but when. As technology advances and regulations catch up, they are poised to become a fixture of everyday life
Now that these AV companies are clocking hundreds of paid trips by the week, “the growth will continue,” Spieczny said. The market is already projected to multiply from its $3.3 billion valuation to $50 billion by 2032. “Once people see for themselves that autonomous vehicles are safe and affordable, they’ll be much more open to hopping in a robotaxi,” he predicted.
Once they are fully integrated into society, robotaxis are expected to transform urban transportation, reducing congestion and providing a more efficient way for commuters to get around. With fleets of self-driving cars zipping through cities, passengers could request rides on demand through an app, eliminating the need for car ownership altogether. And as cities become more automated, these vehicles could seamlessly connect with other smart infrastructure, like traffic lights and public transportation, creating a more cohesive mobility system.
“Everything that moves — at some point in the future — will require autonomous systems,” Spieczny continued. “Cars and trucks are just the beginning of that overall embodied AI future.”
Major Developments in Robotaxi Industry
The evolution of the robotaxi industry is marked by a series of rapid technological advancements, regulatory hurdles and commercial milestones. These are some of the most important events that have transformed autonomous vehicle technology from a futuristic concept to a practical (albeit challenging) reality.
Waymo Announces Expansion into Europe (October 2025)
Following a rapid expansion across the U.S., Google-backed Waymo announced plans to launch its autonomous ride-hailing service in London in 2026. This move could make Waymo the first American company to offer robotaxi services in Europe. However, it could quickly become a crowded market, as Uber and Lyft have announced similar plans to offer autonomous rides in the United Kingdom and Germany through their respective partnerships.
Tesla Begins Paid Robotaxi Service in Austin (June 2025)
In a major step toward commercializing its autonomous driving technology, Tesla launched a paid robotaxi service in Austin, Texas. The initial rollout used a fleet of modified Tesla Model Y vehicles with safety monitors on board. This service represents a significant milestone for the company's long-term strategy, demonstrating a path from its "Full Self-Driving" (FSD) beta program to a revenue-generating transportation service.
Tesla Unveils the “Cybercab” (October 2024)
Tesla unveiled its dedicated, purpose-built robotaxi vehicle at its We, Robot event in 2024. The two-passenger vehicle, referred to as the "Cybercab," featured a design inspired by the Cybertruck and notably lacked a steering wheel or pedals. The unveiling solidified Tesla's vision to create a separate robotaxi fleet rather than relying solely on a network of owner-operated vehicles. However, it likely won’t go into production until 2026 at the earliest.
Cruise and Waymo Receive Driverless Deployment Permits in San Francisco (August 2023)
The California Public Utilities Commission (CPUC) granted Cruise and Waymo authority to expand their driverless robotaxi services in San Francisco. This approval allowed both companies to begin charging fares for rides at any time of day — a significant step toward full commercialization. The decision marked a major regulatory milestone for the industry, allowing for an increase in the number of vehicles and operational hours in a major U.S. city.
Waymo Offers First Driverless Rides to the Public (October 2020)
Waymo became the first company to offer fully driverless (Level 4) robotaxi rides to the public, with no human safety driver on board. The service was initially made available to Waymo One members in a limited area of Phoenix, Arizona. It marked a pivotal moment, transitioning robotaxi technology from a test-and-pilot phase to a commercial service accessible to the public.
Elon Musk Announces the Tesla Network Vision (April 2019)
During the company's Autonomy Investor Day, Tesla CEO Elon Musk outlined his vision for a "Tesla Network" of robotaxis. He stated that owners of FSD-equipped Tesla vehicles would be able to add their cars to a shared fleet, allowing them to earn passive income from autonomous rides. This ambitious plan set Tesla's strategy apart from competitors like Waymo, which were building and operating their own fleets.
Waymo Spins Off from Google (December 2016)
Google’s secretive self-driving car project, which began in 2009, was officially spun out into its own standalone company under the Alphabet Inc. umbrella and renamed Waymo. The move signaled a critical transition from a research "moonshot" to a commercial enterprise with a clear mission. By this time, the project had driven millions of miles on public roads, giving Waymo a significant advantage in its new life as an independent company.
The DARPA Grand Challenge Spawns Modern Autonomy (2004)
The DARPA Grand Challenge competitions, held between 2004 and 2007, significantly advanced autonomous vehicle technology. Stanford University’s winning team, led by entrepreneur Sebastian Thrun, would later form the core of what became Google's self-driving car project, directly linking the military-sponsored competition to the foundational development of modern robotaxi technology.
Frequently Asked Questions
How much would a robotaxi cost?
At roughly $2 per mile, riders have reported that robotaxis cost about the same or less than traditional ride-hailing options like Uber or Lyft.
What does a robotaxi do?
A robotaxi is a fully autonomous vehicle that provides on-demand rides without a human driver, using advanced sensors and AI to navigate city streets and transport passengers from point A to point B.
Are robotaxis legal?
Yes — cities like San Francisco, Phoenix and Beijing allow commercial robotaxi operations, while other places are still in the testing or regulatory approval phase.
