What is a self-driving car?

A sleek, metallic car with a glass rooftop is depicted in a 3D image against a gray backdrop, giving off a futuristic vibe.

Self-driving cars, or autonomous vehicles (AVs), have zoomed around our popular imaginations for decades. In recent years, films and television shows such as WALL-E and Silicon Valley have poked fun at early versions of self-driving cars—and even further back, in the 1980s series Knight Rider, a retro AV was David Hasselhoff’s sassy sidekick.

Get to know and directly engage with senior McKinsey experts on autonomous vehicles

Ani Kelkar is a partner in McKinsey’s Boston office; Johannes Deichmann is a partner in the Stuttgart office; Kersten Heineke is a partner in the Frankfurt office; Martin Kellner is a partner in the Munich office; Philipp Kampshoff is a senior partner in the Houston office; Ruth Heuss is a senior partner in the Berlin office; Takuto Ueha is a partner in the Southern California office; Timo Möller is a partner in the Cologne office; and Ting Wu is a partner in the Shenzhen office.

But self-driving cars are no longer the stuff of science fiction. The software company Waymo (an Alphabet company) currently offers ride-hailing services with its AVs in Los Angeles, Phoenix, Austin, and San Francisco. In China, autonomous taxis and shuttles are even more normalized. Apollo Go, a unit of the Chinese tech company Baidu, has more than 400 robo-taxis on the roads in Wuhan. Other companies are launching public trials in various cities across China and the United States, as well as in Oslo, several German cities, the United Arab Emirates, and Singapore.

At its most basic level, an autonomous vehicle is one that uses advanced software, hardware, and services to operate by itself, with minimal human intervention. Self-driving vehicles stand to provide more affordable and convenient transportation options for passengers, freight, and urban mobility. But challenges remain, including those related to safety, economic viability, and regulation.

For more on the state of the self-driving-vehicle industry, read on.

Learn more about McKinsey’s Automotive & Assembly Practice and the McKinsey Center for Future Mobility.

What is a Level 4 AV?

With the advent of self-driving cars, all vehicles now exist on a five-level scale. Here are the levels:

Level 0 (L0) vehicles have no automation, and a human driver performs all tasks.
Level 1 (L1) vehicles have driver assistance capabilities, including some automated features such as cruise control. A driver must still monitor and control the vehicle.
Level 2 (L2) vehicles have partial automation. The vehicle can assist with tasks like braking, accelerating, and steering. A driver must still monitor the environment and be ready to take control of the vehicle.
Level 3 (L3) vehicles have conditional automation, meaning that the vehicle can operate independently under certain conditions (traffic jams, for example) but can request that a driver take over if needed. These cars represent the highest level of automation that’s widely available to today’s consumers.
Level 4 (L4) vehicles have high automation. These vehicles, including driverless taxis, can operate without needing a driver who’s ready to take over. L4 vehicles are currently being tested and deployed in specific environments.
Level 5 (L5) vehicles are fully automated. These vehicles can operate in any environment and under all conditions without a human driver. They’re the final frontier of autonomous-vehicle development.

Where in the world is autonomous driving catching on?

Ride-hailers can already order a robo-taxi in the American cities of San Francisco, Los Angeles, and Phoenix. In China, robo-taxis and robo-shuttles are available in several cities, including Beijing, Shanghai, Shenzhen, and Wuhan. Testing is also underway in many European cities, including Hamburg, Munich, and Oslo.

Yet across these markets, the outlook for autonomous vehicles is variable. Respondents to a 2023 survey of leaders in the autonomous-driving industry, conducted by the McKinsey Center for Future Mobility, predict that three or fewer companies will capture a dominant share of the global AV market. On a more granular level, it’s a different picture. Only 15 percent of respondents expect that the North American market will be dominated by one or two players. By contrast, 38 percent of respondents predict that Europe’s autonomous-vehicle market will be dominated by two or fewer players.

Learn more about McKinsey’s Automotive & Assembly Practice and the McKinsey Center for Future Mobility.

When will self-driving cars be available?

Survey responses also indicate expected delays in the timeline for adoption: by two or three years across all autonomy levels. Respondents now predict that L4 robo-taxis will become commercially available at scale by 2030, while fully autonomous trucking is expected to reach viability between 2028 and 2031. In terms of deployment, respondents were evenly split on whether China or North America would be the first to roll out L4 highway pilots, which indicates that China is making progress with self-driving cars, likely driven by its growing investments in research and data availability, and receptive customer attitudes, among other factors.

What is shared autonomous mobility?

Most people’s first interactions with self-driving technology will be in shared vehicles, such as robo-taxis and robo-shuttles. Shared vehicles can be a way for autonomous-mobility organizations to generate initial traction, especially in urban areas where the economics of their offerings are more favorable.

That’s easier said than done, though. To attract riders, shared autonomous mobility must be either less expensive or more convenient than conventional options for urban transportation. But the price must also be high enough that all businesses along the value chain can profit. Satisfying both objectives can be difficult, especially if other ride-hailing services are available.

As of 2024, McKinsey estimates that the combination of unit costs, city-level costs, and global costs for self-driving cars could amount to $8.20 per vehicle mile traveled in a typical US city with 1,000 vehicles in operation at any given time. Assuming large-scale operations by 2035, that cost could fall to about $1.30 per mile traveled.

Learn more about McKinsey’s Automotive & Assembly Practice and the McKinsey Center for Future Mobility.

What is remote driving?

Remote driving is when an off-site driver controls a vehicle’s braking and steering to navigate real-time road conditions, traffic, and unexpected obstacles, all of which are shared via sophisticated communications systems and video feeds. Remote driving is a mobility innovation that could accelerate and complement autonomy (exhibit).

Remote driving could complement autonomous mobility by extending the operating range of self-driving cars, allowing them to be used in areas where autonomous driving is either forbidden or unfeasible. A variety of company types may benefit from remote-driving services, including organizations that frequently need to relocate vehicles: for example, rental car companies, agricultural organizations, and defense companies. Repair shops could also offer remote-driving services to return cars to customers after they’ve been serviced. Customers might also want to activate remote-driving services for airport pickup and drop-off, or when they’re coming home after a late night out.

The market seems primed for remote-driving services. In a recent McKinsey survey of about 1,500 car owners in China, Germany, and the United States, about 70 percent of premium car owners and 55 percent of midrange car owners each said they would consider using remote-driving services. Globally, car owners said they would be willing to pay around $53 per hour for remote-driving services.

What’s missing from today’s remote-driving market that could support its future growth? McKinsey has identified three key factors:

Customer acceptance. Based on our own survey, it seems that many people are already open to remote-driving services, but it will be important to continue monitoring customer sentiment over time.
Insurance coverage. Regulators and insurers must create clarity about liability for all stakeholders, including the car manufacturer, the remote-driving service provider, and the human driver who operates the vehicle. If insurers decline to create products that apply to remote driving, service providers may need to offer them instead.
Safety studies and regulatory frameworks. As with autonomous driving, regulators will need to carefully scrutinize remote-driving products and services before approving pilots.

What factors will help accelerate AV development?

Before widespread AV adoption is possible, there needs to be more autonomous vehicles. The same survey revealed that increased investments in software development (including prediction algorithms and perception software) are needed for more vehicles to achieve full autonomy. For instance, more than $4 billion is needed for full-journey autonomous trucks, more than $2 billion for Level 3 highway use cases, and more than $5 billion for robo-taxis in Levels 4 and 5.

While consumer demand is not seen as a major impediment to adoption, two-thirds of survey respondents believe improved safety will be important for consumer uptake. Secondary consumer considerations will be greater productivity—that is, the ability to multitask while driving—as well as comfort.

When it comes to achieving profitability, strategic partnerships will be key. Almost all respondents to our survey—96 percent—say partnerships are crucial to the development of autonomous vehicles. Collaboration among stakeholders will be critical to derisking investments and creating the infrastructure needed to build, operate, and maintain autonomous vehicles at scale. What’s more, partnerships can help drive further innovation.

What will AV development mean for the future of freight?

Autonomous-mobility vehicles stand to significantly transform the future of freight transportation by enhancing efficiency, reducing costs, and addressing labor shortages. More specifically, autonomous freight technology can optimize routes, improve fuel efficiency, and minimize downtime through predictive maintenance, all of which contribute to lower operational costs. These efficiencies are critical in an industry where margins are tight and the competition is fierce.

Self-driving trucks could also help mitigate the shortage of truck drivers, which is especially acute in Europe. The trucking industry has long faced challenges in attracting and retaining drivers. But by integrating autonomous trucks into their fleets, trucking companies could alleviate some of their reliance on human drivers and the accompanying hiring-related pressures. Although fully autonomous trucks may not completely replace human drivers in the near term, they can handle long-haul segments of a journey while leaving the more complex urban driving to human operators. This hybrid model can extend both the reach and efficiency of freight operations.

Circular, white maze filled with white semicircles.

Finally, the adoption of AV technology in freight transport is likely to spur innovation and new business models within the industry. For instance, the increased efficiency and lower costs associated with autonomous trucks could lead to the development of new logistics solutions and services. Companies may also explore more dynamic and responsive supply chain strategies, such as just-in-time delivery and on-demand freight services. What’s more, the shift to AVs could encourage investments in supporting infrastructure, such as dedicated lanes for autonomous trucks and smart logistics hubs, which would further enhance the overall mobility ecosystem.

How could autonomous mobility affect jobs?

Autonomous mobility is poised to have a significant effect on jobs across various sectors. Adoption of autonomous vehicles will likely lead to a reduction in demand for certain roles, particularly those involving routine driving tasks. But the transition won’t happen overnight, and there will be a period where human oversight and intervention will still be required, especially in complex driving scenarios.

On the flip side, the rise of autonomous mobility is also expected to create new job opportunities. The development, maintenance, and management of autonomous-mobility technology will require specialized skills in areas such as software development, data analysis, and specialized system maintenance. There will also likely be new roles in remote monitoring and control of autonomous fleets, as well as in customer service and support for autonomous-mobility-related services.

Learn more about McKinsey’s Automotive & Assembly Practice and the McKinsey Center for Future Mobility. And check out jobs related to autonomous mobility if you’re interested in working with McKinsey.

Articles referenced: