Autonomous driving steps out of "science fiction": shared mobility enters the large-scale speed race.
For the autonomous driving industry, ten years ago, the focus of debate in the industry was "whether it is possible to build an autonomous vehicle"; today, the topic has shifted to "how to roll out shared autonomous driving services in a safe and profitable manner". As unmanned driving commercial services are gradually implemented and the supply chain system matures, a brand-new collaborative relationship has taken shape between automakers, autonomous driving technology suppliers, fleet operators and mobility platforms. "Scaling" has replaced "technical feasibility" and become the core keyword for the development of the industry.
Market Size: Expected to Reach USD 430 Billion in 2035
The shared autonomous driving market model built by McKinsey based on both supply and demand sides predicts that by 2035, the global shared autonomous driving market (including Robotaxis and autonomous driving shuttles) will reach a size of USD 190 billion to 430 billion, corresponding to an operating fleet size ranging from 1 million to 2.5 million vehicles.
From a regional perspective, China and the United States will be the two core markets in the world. By 2035, the U.S. market is expected to reach USD 100 billion to 190 billion, corresponding to 400,000 to 800,000 operating vehicles; the Chinese market will be around USD 60 billion to 150 billion, with an operating vehicle size of 400,000 to 1.2 million. The difference in market size between China and the United States mainly comes from the expected gap in cost per mile — the cost per kilometer in China is expected to be about 50% lower, which essentially reflects the difference in purchasing power levels between the two regions.
The European market will be about USD 20 billion to 70 billion, corresponding to 100,000 to 400,000 vehicles; the Middle East market will be about USD 10 billion to 20 billion, with about 60,000 to 100,000 operating vehicles. It is worth noting that the implementation form in Europe may be different from that in China and the United States: in the future, the European market will most likely be dominated by robot shuttles, with in-depth cooperation between technology suppliers and local public transport operators. Such services can not only act as branch supplements to public transport, but also apply to areas where the expansion cost of traditional public transport is too high and the implementation cycle is too long.
Implementation Pace: The United States Leads in Commercial Expansion
In terms of implementation pace, the United States takes the global lead in commercial advancement. At the end of 2024, only 3 cities in the United States provided commercial autonomous driving services; by September 2026, the number of covered cities has increased to 15. If the current expansion pace continues, by the end of 2027, the number of cities with commercial autonomous driving services in the United States will reach about 40. In Europe, the first batch of commercial services have been launched in cities such as London and Zagreb, and more cities will follow up in 2027; China has already had multiple urban routes operating on a regular basis, and the Middle East has also launched services in cities including Abu Dhabi and Riyadh.
The key regulatory constraints are also gradually loosening. In June 2026, the United Nations Economic Commission for Europe (UNECE) passed the world's first regulatory framework for fully autonomous driving systems, aiming to provide a reference for countries to formulate unified safety standards. Although this framework cannot immediately solve the problem of fragmented regulations in various countries, it has provided an alignable benchmark for the industry.
Consumer Attitudes: From Novelty Experience to Usage Habit
To investigate consumers' acceptance of autonomous driving mobility, McKinsey's 2026 *Mobility Consumer Pulse* survey covered more than 20,000 mobility users in China, Germany, Japan, the United Kingdom and the United States, and the conclusion is more optimistic than the industry expected: among users who have experienced shared autonomous driving services, 87% of respondents gave positive feedback, and their trust in autonomous driving increased significantly after the first ride; among respondents who have never used such services, 43% said they are willing to try in the future, and this proportion is about 10 percentage points higher among the group that "knows such services are already operating locally". Users' familiarity with the service is forming a "habitualization effect", and the proportion of skeptical respondents is less than 20%.
The core support for the improvement of user acceptance comes from four factors: safety, accessibility, sustainability and cost-effectiveness. In terms of safety, the 2025 operating city data disclosed by Waymo shows that under the same driving mileage, the injury accident rate of autonomous vehicles is 82% lower than the average level of human driving; in terms of sustainability, under the pure electric mode, the full life cycle carbon emission of shared autonomous vehicles per mile is 85% to 98% lower than that of private diesel cars; in terms of accessibility, the service provides new mobility options for people who cannot afford to buy a car, or have limited travel due to age and physical conditions. However, cost is still the biggest pain point at present: 37% of existing users list it as their main dissatisfaction, and users who have not experienced such services are more price-sensitive.
Cost Tipping Point: Cost Per Mile Drops by Nearly 80%
Whether the industry can cross the core threshold of scaling ultimately depends on the unit operating economy. McKinsey predicts that by 2035, the cost per mile of shared autonomous vehicles will drop by nearly 80%, and the driving forces for cost reduction come from three directions: the cost reduction of autonomous driving technology itself, the improvement of vehicle service life and operation cost-effectiveness, and the scale effect formed by fleet operation. At present, some autonomous driving enterprises have begun to disclose the break-even status on a vehicle basis in specific markets.
It should be noted that there is still high uncertainty in the actual development of shared autonomous driving. The global regulatory system remains fragmented, which has slowed down the industry implementation pace for many times in history; operational bottlenecks such as charging facilities, safety operator configuration and dedicated station demands have not been systematically solved; the industry still requires high-intensity capital investment, and the large-scale verification of unit economy has not been completed; the wide replacement of services may also have an impact on related employment, and the differences between regions are very significant.
Value Chain Disassembly, Focus on Core Competencies
As the industry ecosystem is gradually layered, the development path of full-link vertical integration is being gradually abandoned, and all parties in the industrial chain begin to return to their own core competencies: automakers focus on vehicle integration, technology companies focus on virtual driver technology and safety system demonstration, and fleet operators focus on daily operation and user experience optimization. This specialized division of labor not only accelerates the speed of large-scale implementation, but also reduces the overall capital intensity of the industry.
For automakers, the core opportunity lies in realizing the industrialization of autonomous driving hardware integration: providing efficient vehicle access solutions for safety core components such as sensors, redundant braking and steering, and designing dedicated vehicle models for high-utilization operation scenarios. The requirements of such vehicles in terms of durability, maintenance cost, predictive maintenance, modular battery and fast charging capability are far more complex than those of private cars. In the long run, the revenue model of automakers is expected to shift from "selling vehicles" to "selling availability", and revenue will rely more on fleet leasing, operation duration guarantee and maintenance service packages.
Shared autonomous driving has completed the leap from concept to commercial service, and is gradually entering the daily mobility scenarios of more cities. For enterprises in the industrial chain, the real problem is no longer whether the technology can be implemented, but whether they can create practical and sustainable value in the urban scenarios they have laid out. The window for scaling has opened, but who can balance safety, cost and operation at the same time, the answer is still to be revealed by the market.
This article is from the WeChat Official Account "Auto Market Insight", written by Zheng Li, and published with authorization from 36Kr.