TomTom CPO: Industry shifting from software-defined to AI-defined vehicles

The article notes that not long ago, a new car's story could be told through horsepower, torque and exhaust notes. For next-generation platforms, differentiation increasingly comes from driving assistance, electric range, the digital experience and connected features. The question now shifts from whether a vehicle is software-defined to whether it can understand the road, the driver and the situation well enough to create a more confident driving experience.

The software-defined shift laid the foundations for this moment. Centralized compute makes it easier to run advanced models, and continuous connectivity means vehicles can improve after leaving the factory. AI-defined capabilities now touch the full in-vehicle stack, including navigation, infotainment, battery and energy management, and HVAC, the CPO said.

He highlighted the importance of fusing sensor perception with road intelligence. A vehicle relying only on sensors, he said, can end up driving like a tourist — cautious and reactive. A vehicle that combines sensors with road intelligence behaves more like a local, knowing where lanes split, where speed limits change, and where closures or hazards are likely to appear. This supports smoother behavior, fewer sudden corrections and a more comfortable experience.

Examples of the shift include the move from simple voice assistants to agentic AI-powered assistants that handle complex, multi-faceted requests across the car's systems. End-to-end perception models can also pick up subtle cues, such as when the vehicle ahead is driving erratically, and adjust accordingly by creating more space.

According to the CPO, the bottleneck moves from whether we can build a clever model to whether we can feed it with reliable, high-quality context at scale. Mapping, routing and road context give a vehicle the understanding it needs to look beyond the immediate moment and translate raw signals into safe, assured behavior.