The future of mobility is moving fast, but not all manufacturers are following the same path. On the one hand, Stellantis has decided to pull the brake on his plan to autonomous driving, on the other Volkswagen pushes on the accelerator with the help of Bosch and his software house LoveTwo opposing strategies, which clearly illustrate the challenges of a fascinating but complex sector, where Tesla and the Chinese brands are leading the race towards increasingly intelligent and software-controlled cars.
Stellantis freezes AutoDrive: the Level 3 dream postponed
Secondo ReutersStellantis' decision stems from a combination of factors: Costs high levels of development, complexity technical and question market still weak. AutoDrive, the Level 3 autonomous driving system presented in February as “ready”, which would have allowed drivers to take their hands off the wheel and their eyes off the road under certain conditions, will therefore remain on the shelf, at least for now.
The group confirmed that the technology is “developed and available,” but there are no conditions for an immediate commercial launch. This freeze is in addition to other rethinks of the group's digital strategy: the partnership with Amazon for the SmartCockpit, while the Italian-French giant will increasingly rely on Android solutions and external suppliers for more complex software functions.
The CEO Antonio Filosa It wants to focus internal efforts on the distinctive elements of the customer experience, leaving the most technologically expensive and complex areas to specialized companies.
aiMotive: Stellantis's hidden secret
The future of autonomous driving at Stellantis could be aiMotive, the Hungarian startup acquired in 2022 and specializing in AI and advanced driver assistance systems (ADAS). It remains to be seen if and when a new version of AutoDrive will arrive and whether it will still include Level 3 features.
The context isn't helping: the group formed from the merger between FCA and Peugeot is emerging from a difficult 2024, with declining sales and a stock price down more than 40% in a year, closing at $9,93. Filosa will have to present a clearer software strategy by 2026 to support its goal of €20 billion in annual software revenue by 2030.
Volkswagen and Bosch: Turbocharging AI and Autonomous Driving
While Stellantis slows down, Bosch e Love, the group's software development division Volkswagen they step on the accelerator. With their Automated Driving Alliance, are developing an AI-based software stack that promises to bring Level 2 and 3 assisted and autonomous driving systems to a safer, more scalable, and more accessible standard. The German giant will integrate these functions into the new software-defined architecture of its vehicles, starting from 2026. Not only that: Bosch will also make the solution scalable for other manufacturers.
“We are demonstrating that German industry masters the key technologies of AI and autonomous driving,” he explains. Peter Bosch, CEO of Cariad. For Mathias Pillin, CTO of Bosch Mobility, “data and artificial intelligence are fundamental: only by working together can we bring reliable systems to the road on a large scale.”
End-to-end AI: From sensors to decisions
The new Bosch–Cariad stack is not limited to a single component, but applies theartificial intelligence along the entire technology chain: from object recognition to the fusion of data from cameras and radar, up to decision-making and safe control of the engine, steering, and brakes.
The next step will be the integration of systems Vision-Language-Action (VLA), capable of connecting images and language to "think" like an expert driver. This will allow the software to anticipate complex urban traffic scenarios and recognize hidden dangers, an approach reminiscent of the most advanced generative AI.
What makes the difference is the approach data-drivenThe test fleets, which already include models like the ID.Buzz and the Audi Q8, are updated several times a day with new software versions. Tests are taking place in Europe, Japan, and the United States, and hundreds of new vehicles equipped with advanced sensors will enrich the data collection throughout 2025. The goal is to train artificial intelligence on so-called corner cases – those rare and unpredictable situations that truly put an autonomous car to the test – to bring increasingly reliable, safe, and comfortable systems to series production.
