Here’s one reason why GM may have ditched Cruise robotaxis: carmakers will earn more than half of their revenue from software and digital services by 2035.
That’s according to a report from IBM, Alliance for Automotive Innovation, AWS, and Red Hat, which surveyed 1,200 industry executives to predict what’s next for cars. In short: electric, software designed and some, but not full, autonomy.
Three-quarters of respondents predicted that vehicles will be “software defined” and powered by AI.
A software-defined vehicle (SDV) is when everything about the car is controlled by software, not just the entertainment or navigation, but how the car works, its features, and even how it drives — and that includes driverless technologies, with two-thirds of industry execs believing customers will expect autonomous features of some sort by 2035. Potentially, SDVs can be upgraded via their software, sort of like when Apple releases an iOS update that adds tools to an iPhone.
“They foresee an extensive transformation where software and AI are applied not just to in-car experiences but to the very core of the vehicle — its controls, its functions, and its interactions with the driver and the world around it,” the report said.
Driverless dollars now
And car execs expect to make 51% of their revenue across such digital services and car software by 2035, up from 15% today. The shift to a subscription model could boost margins in the industry — and that means higher costs for customers.
Indeed, IBM predicted that autonomous driving features will help generate $269 in monthly recurring revenue by 2035.
Some of that will come from “over the air” (OTA) upgrades, with car owners able to upgrade features in order to extend the life of the vehicle or take advantage of new services. The survey showed that six in ten execs believe customers will want this function by 2035.
Not driverless, but some autonomy
No wonder carmakers are stepping away from robotaxis — GM pulled funding for autonomous taxi firm Cruise, in favour of investing in autonomy for private vehicles.
The report noted that 18% of carmakers plan to invest to develop in-house capabilities for developing vehicle technology systems, including automation.
Of course, being able to extract extra cash from car owners doesn’t mean there’s no money to be made in robotaxis — after all, Waymo just expanded to Tokyo, its first foreign market. But the reports findings do point to an easier, faster route to paying for driverless development by adding autonomy to cars for a fee, just as Telsa has long done.
Indeed, Tesla charges $99 for what it calls supervised full-self driving, though that requires an eligible vehicle. In the US, Ford charges $49.99 a month for BlueCruise, its version of level-2 driverless autonomy, but again only on eligible vehicles; there’s also an option for a one-off payment of $2,495.
Slow road to driverless and SDVs
While the report predicts two-thirds of car drivers will expect some autonomous features by 2035, they believe those features will stay limited to level two and level three — levels of autonomy where the driver must maintain attention and be ready to take back the wheel. The execs predicted fewer than a quarter of the market will have more advanced features by then.
“Even if the technical challenges for high-level autonomous driving are cleared, regulatory issues and societal acceptance take more time and effort,” the report quoted Kenichi Takagi, Vice President, Connected Systems R&D, Denso International America, as saying.
Autonomy and SDVs both face challenges. One is security — after all, a hacked car is a scary thought indeed — while another is skills, as carmakers need staff with strong software talents alongside knowledge in vehicle development. The biggest challenge, according to the study, was complexity, with a need for standard interfaces between technology layers to allow in-car systems to safely share information.
“Automakers are increasingly moving to highly centralized, powerful high-performance compute (HPC), or domain control, units that can separate hardware and make software more manageable and easier to update,” the report noted. “But this shift is exposing issues, with 79% of executives citing the technical complexity of separating the hardware and software layers as a moderate or significant SDV challenge. In fact, 47% say it is their number-one challenge.”
And that means that SDVs may have a hard road ahead to arrive in force by 2035, regardless of what execs say today.
