By Manuel Nau, Lead Editor at IoT Business News.
Berg Insight forecast that Global Robotaxi Fleet will expand from around 6,500 vehicles by the end of 2025 to 1.53 million by 2036, as Level 4 autonomous driving platforms move from limited deployment to larger commercial fleets.
Scaling autonomous driving becomes less of a question of whether the vehicle can navigate without a human driver and more of a question of whether all the systems around the vehicle can replicate the economy in the city. Robotaxis combines sensors, onboard computing, vehicle control systems, connectivity, mapping, fleet infrastructure and remote operations, making commercial expansion more complex than simply using additional vehicles.
Against this background, Berg Insight expects the global robotaxi fleet to reach 1.53 million vehicles in 2036, compared to approximately 6,500 operating worldwide by the end of 2025. 79.2 percent.
Driverless robotaxi services are currently concentrated mainly in the United States, China and the Middle East, although services have recently emerged in Europe, South Korea and Singapore. Geographic expansion is more desirable as technology providers, mobility platforms and fleet operators try to produce service models that work in additional markets.
The technology provider is the integration layer
One of the aspects that differentiates the robotaxi market from the more conventional vehicle business is the incredible amount of technology integration that is concentrated in automated driving providers. The company does not provide only one subsystem. They developed Level 4 automated driving systems and determined how perception software, sensors, computer hardware and vehicle controls work together.
Berg Insight identified Avride, Baidu through Apollo Go, May Mobility, Mobileye, Momenta, Motional, Pony.ai, Tesla, Waymo, Wayve, WeRide and Zoox among the leading robotaxi technology providers. Some of these companies also control supporting functions including mapping, simulation, validation, fleet data infrastructure and remote assistance.
Martin Cederqvist, Senior Analyst at Berg Insight, said:
“Automated driving systems and the data used to train and validate those systems are key proprietary assets.”
This structure has important operational implications. As the fleet grows, the vehicle becomes just one component of a more distributed computing system. Data generated in the field must support the validation and improvement of driving systems, while fleet operations may rely on remote assistance and centralized infrastructure. For technology providers serving the connected vehicle market, robotaxis therefore integrates onboard intelligence with continuous fleet-level data operations rather than treating connectivity as an isolated vehicle feature.
Commercial scale depends on more than autonomous driving
Mobility platforms form another important layer of the evolving ecosystem. Companies including Uber, Lyft, Bolt, DiDi, CaoCao Mobility and T3 Mobility can provide applications, orders, payments, pricing and customer support while connecting autonomous vehicles with existing demand.
The distinction is important because a technically capable vehicle does not automatically create a commercial service. Robotaxi’s economy depends on maintaining the vehicle adequately while covering vehicle, technology and operational costs. Established mobility platforms can contribute customer base and data on travel patterns, which can help operators match autonomous fleet capacity with demand.
A more difficult problem is replicating the deployment across all locations. Different traffic environments, regulations and operating conditions can create additional engineering and data requirements when services enter new cities. Berg Insight points to end-to-end AI and large-scale driving models as technologies that can reduce reliance on manually defined driving rules and increase generalization across driving environments.
This is one of the most important changes in robotaxi architecture. If the driving system can reuse the behavior learned in the market, the expansion becomes less dependent on extensive city-specific techniques. However, Berg Insight notes that the amount of additional local data collection still varies between markets and systems, meaning that geographic scalability will continue to depend on the automated driving approach and deployment environment.
The design of the vehicle will also affect the economy
Factory-integrated and purpose-built robotaxis can change the cost structure if production volumes become large enough. Integrating autonomous vehicle hardware and systems during manufacturing can reduce some of the complexities associated with adapting conventional vehicles for driverless operation.
For OEMs, system integrators and connectivity providers, the forecast points to a market where differentiation will increase due to the ability to operate a complete fleet of vehicles instead of just individual components. Autonomous driving software, vehicle architecture, fleet data systems, remote operation and passenger platforms must all work as a coordinated service.
Berg Insight forecasts more than 1.5 million robotaxis by 2036, thus representing more than expected in the number of autonomous vehicles. Reaching that scale requires the industry to transform its integrated location-specific deployment into a repeatable operational platform – while meeting security requirements, local regulations and the economics of commercial mobility services.