McKinsey Report Robotaxis will be cheaper than your own car in 2035

Related Vendors

FAULHABER Drive Systems

dataTec AG

Wuxi InfiMotion Technology Co., Ltd.

Robotaxis at 80 cents per kilometer, app store models for cars, and $84 billion in supplier potential—McKinsey reveals the radical transformation of the automotive industry.

At CES in Las Vegas, McKinsey shared key research findings on autonomous mobility and software-defined vehicles, offering insights into the current state and future direction of the automotive industry.

Since 2010, more than $850 billion has been invested in four key areas: automated and connected mobility, electromobility, and shared mobility. According to Phillip Kampshoff, Senior Partner at McKinsey, 93% of this funding has come from outside the automotive industry. “This is the reason for the extremely fast pace of innovation in recent years.”

Autonomous Driving: Back from the Downturn

Autonomous driving experienced a boom in 2015 before losing momentum as the focus shifted to electromobility. Despite setbacks (e.g., Argo AI, Cruise), the market has been growing strongly again since 2024. This is reflected in the valuation of companies, explains Kampshoff.
 When will we see the large-scale introduction of autonomous vehicles? “First, it must be affordable for consumers, but at the same time, it also needs to be profitable for service providers,” says Kampshoff.

For robotaxis, the consultant sees significant potential for cost savings. Currently, the most expensive part of a taxi service is the driver, accounting for around 60% of costs. Once autonomous systems replace human drivers, operating costs will decrease significantly.

"Robotaxis will be worthwhile for 50% of people in a major city"

According to studies, in a city like Washington, D.C. (approx. 690,000 residents), owning a personal car becomes more cost-effective than using ride-hailing services with drivers if an individual drives more than 2,000 miles (3,200 km) per year. However, with autonomous taxis, this threshold would shift to 7,500 miles (approx. 12,000 km) per year. This means that, from a purely economic perspective, car ownership would no longer be viable for 50% of people in a major city—provided that robotaxi operating costs continue to decline.

Today, the cost of a robotaxi ride is around $8 per mile (€4.85/km), according to a recent McKinsey study cited by Kampshoff. By 2035, this could drop to $1.32 per mile (€0.80/km), provided that costs for sensors, operations, and maintenance are reduced.

Strong Acceptance of Robotaxis in China

Another key factor is public acceptance. In the U.S., only 11% of people would currently use a robotaxi (Germany: 11%). In contrast, the figure in China is 31%. Customers there are much more open to vehicle connectivity, explains Kampshoff, whereas in the U.S., safety concerns dominate. However, initial U.S. data does not support these concerns: an autonomous vehicle drives 820,000 miles (1.3 million km) before being involved in an accident with injuries or fatalities, compared to human-driven vehicles, which average about 350,000 miles (560,000 km) before such incidents. "This suggests that autonomous vehicles are 2.3 times safer," Kampshoff notes, while acknowledging that the current number of robotaxis is still relatively low.

Cities like Phoenix, San Francisco, and Los Angeles already have autonomous vehicles operating on a large scale. "At this point, we are beyond the pilot stage," says Kampshoff. "We are in the early phase of deployment and scaling." However, major challenges remain, especially in regions with extreme weather conditions like snow. In Germany and the EU, the rollout of autonomous vehicles is expected to be slower due to complex regulatory hurdles and stringent safety requirements, Kampshoff predicts.

Manufacturers Know How to Sell Cars—Not Software Services

The automotive industry is undergoing its next major transformation: the shift to software-defined vehicles. This transition allows manufacturers to accelerate innovation cycles and strengthen customer relationships beyond the initial purchase, says Eduardo Mañas Pont, Senior Engagement Manager at McKinsey.
 He compares it to an iPhone: when a customer buys one, they enter a long-term relationship with Apple by making purchases in the App Store. "This is an experience that the automotive and many other hardware industries currently lack because they sell a product limited to a single transaction," Mañas Pont explains.
 To decouple hardware development cycles from software cycles, manufacturers must simplify their architecture by transitioning from numerous control units to a centralized structure. Once a successful shift to a software-defined architecture is achieved, the key question becomes: What do consumers want? Which comfort or assistance features are they willing to pay for?

Suppliers: Software Potential Exceeds $80 Billion

Established companies, in particular, are struggling with this transformation. In contrast, new players that have been software-centric from the start are often far more productive. That’s why partnerships are a key strategy, says Mañas Pont.
 The automotive value chain is also evolving, especially from the perspective of suppliers. "You're taking a car that was primarily composed of mechanical components and turning it into what is essentially a smartphone on wheels," explains Mañas Pont. Many mechanical components are becoming commoditized or even eliminated from vehicles. This shift presents suppliers with opportunities worth $84 billion in the automotive software sector.
 Software development is becoming increasingly complex, but advances in AI are making the transition easier. "The real challenge is that there is still no industry standard," says Mañas Pont. OEMs rely on suppliers for development, but they also risk backing the wrong approach.