Shared autonomous vehicles (AVs) now operate in more than ten cities worldwide, including Beijing, Oslo, Phoenix, and San Francisco. More deployments are planned in other regions as AVs with level four capabilities—those that can handle most functions without human intervention—become increasingly sophisticated.
While engineers have overcome many technological hurdles for autonomous mobility and continue to make strides, industry leaders should also address other issues—especially those related to the economics of these offerings—to help shared AVs generate initial traction. One major goal should involve increasing uptake in urban areas because the economics in other regions are less attractive. We reviewed the affordability and pricing of shared AVs, as well as concerns related to safety, accessibility, and sustainability, to determine how the market might evolve.
A dual mission: Affordability and profitability
It’s a balancing act. Shared autonomous mobility must be less expensive or more convenient than conventional urban transportation options to attract riders, but the price must be high enough that all businesses along the value chain can profit. Satisfying both objectives simultaneously can be difficult, especially if other ride hailing services are available.
To determine how pricing for shared AV services might evolve, we first examined costs in three separate areas: unit costs, city-level costs, and global costs. Combined, these costs could amount to $8.20 per vehicle mile traveled (VMT) for a typical city in the United States with 1,000 vehicles in operation at a given time. But when we analyzed how total costs might evolve, we found that they could potentially fall to about $1.30 by 2035, assuming large-scale operations (Exhibit 1).
Factors contributing to the improvement are detailed below.
Vehicle
This category includes all costs and depreciation for a single self-driving system with level four capabilities. These costs account for about 20 percent of total costs today, or $1.64 per VMT, but they could decline by about 85 percent by 2035. The major factors behind the decrease include the following:
Greater vehicle utilization. The greatest cost reductions are expected from increased use of shared AVs with level four capabilities. The uptick will largely result from greater consumer adoption, increased utilization, wider operating areas, and improved dispatching algorithms.
AV technology. Companies will likely manufacture some system components, including sensors and high-performance computers, at greater scale, which will lower costs. Other gains could come as engineers simplify the hardware and software for AV kits. For instance, the kits could contain fewer sensors and have a lower storage and compute capacity. Software and data needs will likely be less complex, and engineers could optimize both algorithm performance and software update processes to minimize maintenance and repair needs. Despite their lower costs, the simplified AV kits should maintain or improve performance.
Purpose-built vehicles. These vehicles, as well as their AV kits, would have a longer lifetime. To create them, companies would focus on designs that reduce wear or downtime. They could also improve system integration as volumes rise or use replaceable batteries that are optimized for each use case.
Local operations
One-time and running costs for location-specific operations include those for operation-design-domain mapping, algorithm localization, and validation efforts. Providers should also budget for fleet management, charging infrastructure, launch management and preparation, and other infrastructure-related expenses. Local operations now account for about 50 percent of the total cost of shared autonomous mobility, or about $4.14 per VMT.
If companies increase their fleet size, they could benefit from economies of scale that reduce their local operational costs. In addition, they may be able to reduce local operational costs by about 70 percent by 2035 through improvement initiatives. Improvements in daily operations are expected to deliver the most benefits, and important levers might include automating charging and cleaning procedures, standardizing sensor maintenance and repair processes, creating partnerships with existing mobility and infrastructure operators to share hub facilities, and reducing “empty” miles by optimizing the number and location of hubs and by deploying vehicles based on data. Providers might also consider establishing or increasing access to rideshare networks and platforms to maximize customer convenience and rider volumes.
Beyond daily operations, providers may also decrease local costs through better R&D and launch management. For instance, they might create a specialized team composed of employees who have successfully managed launches in multiple locations to oversee future launches. They could also increase the efficiency of data collection, performance simulations, and other R&D tasks.
Global deployment
Global deployment costs, such as those for launch management teams, now represent about 30 percent of the total cost of shared autonomous mobility, or $2.40 per VMT, but they could potentially decline by about 85 percent by 2035.
As with local operations, improvements in AV technology account for much of the cost reduction. Companies may decrease central data and compute needs as their scale and efficiency rise, and they may also lower hardware expenses through cost-optimized sourcing. Within daily operations, providers may reduce costs by improving algorithms, optimizing vehicle-control-center processes, and enhancing operator interfaces—actions that will minimize the need for vehicle-control-center operators to intervene during rides.
Currently, most shared AV providers operate on a relatively small scale and cover a limited area. As they grow in scale and expand their coverage area, they might benefit from centralizing certain functions, including finance, procurement, HR, and legal, which may be less costly than having independent groups at individual locations. Companies may also explore options for automating or off- or nearshoring some central tasks.
Safety: Improving systems while building consumer confidence
Surrendering vehicle control to an autonomous system requires great trust in the underlying technology, and many people might be understandably wary the first time a driverless vehicle arrives at their door. Exacerbating their concerns, the media has recently reported on multiple incidents in which AVs behaved erratically in traffic, were involved in accidents, or injured pedestrians. Regulators are closely watching developments, and reporting requirements may become more stringent across the board.
As shared AV operators try to scale their operations, they will need to address lingering safety concerns. In a 2024 McKinsey consumer survey, 53 percent of respondents stated that safety concerns were a major roadblock to more widespread AV adoption (Exhibit 2). The need for better road infrastructure to support AVs was a distant second at 35 percent. Although safety issues still loom large, the percentage of respondents citing it as a concern was lower than the 56 percent reported in the 2022 survey.
Companies are continuing to improve safety by perfecting their AV hardware and software through improved algorithms and greater testing and validation. Their leaders are also increasingly viewing safety as more than a technical issue, and some are creating dedicated safety organizations and establishing frameworks with third parties to improve safety standards.
Better governance and reporting mechanisms could also build trust in the safety of AVs. For instance, companies could frequently communicate with policymakers about important developments or work with regulators to establish standard frameworks for monitoring and tracking the safety of self-driving systems. To allay consumer concerns, companies may want to educate the public about ongoing advances in autonomous driving, as well as the social, economic, and environmental advantages of AVs.
Accessibility: Ensuring that certain groups can use AVs
As shared AV ridership grows, operators must guarantee equitable access for populations that might have limited driving abilities. Some early leaders are already minimizing obstacles for older adults and people with disabilities through the following:
Barrier-free vehicle designs. Providers ensure that some vehicles in their fleet are wheelchair accessible to allow riders more independence.
Partnerships. Community organizations that advocate for people with disabilities have worked with shared-AV operators to help shape and test accessible ride hailing services.
With such accommodations, using a shared AV could become just as convenient as owning a private vehicle or hailing a taxi. For older individuals and those with disabilities, the service could provide greater independence and decrease isolation by allowing them to leave their homes more often.
For underserved groups with limited transportation options, shared AVs might provide another affordable transportation option if prices continue trending down. For instance, they might operate in areas where human rideshare drivers tend to decline rides because of safety concerns or other issues. Adding incremental routes for AVs is less expensive than for driver-based services, allowing providers to scale up their operations more quickly.
Sustainability: Even further emissions reductions
Most autonomous fleets contain battery electric vehicles (BEVs). Shared BEVs produce 85 to 98 percent fewer emissions per passenger mile traveled than private diesel vehicles. A recent McKinsey analysis suggests that companies could reduce current BEV emissions by about 71 percent by making improvements throughout the entire vehicle life cycle, from the design phase through operations (Exhibit 3). Supply chain improvements, such as using green components and energy, could account for about 42 percent of the decrease (or 11 grams of CO2 per passenger kilometer).
One caveat: Although shared AVs may rely on electric batteries, their increased use could exacerbate road congestion and raise the number of “deadhead” miles—those with no passenger on board. To ensure that shared fleets capture maximum sustainability benefits, regulators and industry leaders must acknowledge and address these potential complications.
Although shared AVs are now a reality, they still draw a second look, even in cities where they are well established. For these vehicles to gain scale and expand into new locations, stakeholders cannot just strive for technological improvements; they must also focus on affordability, safety, accessibility, and sustainability while ensuring that all businesses along the value chain can profit. Many OEMs and others are already striving to improve these dimensions, and their efforts could make shared AVs an essential component of the urban mobility mix.