Long dominated by US giants (Waymo) and Chinese players (Baidu, Pony.ai), the race for 100% autonomous robotaxis has reached a major milestone on the European continent. In Zagreb, Croatia, the startup Verne — backed by Rimac Group, Kia, and the Saudi Public Investment Fund (PIF) — officially launched the first commercial robotaxi rides operating without any human safety driver on public roads. Ahead of planned trials in Munich or London, Croatia becomes Europe's first commercial testing ground for Level 4/5 autonomous mobility open to the public.
This deployment relies on an international ecosystem and a new decoupling of the automotive value chain: Arcfox Alpha T5 electric vehicles (BAIC), Pony.ai's 7th-generation AI driving stack (with over 100 million km driven), and distribution via Uber and the dedicated Verne app. This milestone confirms the irreversible shift toward the Software-Defined Vehicle (SDV).
SDV Architecture & Embedded AI: The tech shift initiated in Zagreb and Europe
The deployment success in Zagreb proves that a vehicle's core value no longer resides in its mechanical engine or chassis, but in its software architecture. Automakers and software vendors are shifting away from legacy proprietary Electronic Control Units (ECUs) toward modular architectures driven by real-time AI decision models connected to distributed Cloud infrastructure.
For IT directors (CIOs) and software engineers across Europe, North Africa, and emerging markets, this transition requires adapting IT infrastructures to orchestrate the complete data lifecycle.
Tech Deep Dive: 3 software architecture pillars for Mobility IT Leaders
To successfully manage the transition toward connected and intelligent fleets, three key technical priorities emerge for software architects and IT leaders:
1. Shift toward open API architectures and microservices
- AI Stack Interoperability: Hardware platforms must connect via standardized Cloud APIs to integrate any autonomous driving AI stack (Pony.ai, Mobileye, Wayve).
- Third-Party Ecosystem Integration: Fleet management platforms must communicate in real time with major ride-hailing APIs (Uber) and municipal MaaS platforms.
- CI/CD Pipelines & Over-The-Air (OTA) Updates: Deploy ultra-secure continuous deployment pipelines to push algorithm updates and security patches remotely.
2. Establish real-time data governance (Edge-to-Cloud)
- Hybrid Edge/Cloud Processing: Time-critical execution occurs inside the vehicle (Edge computing), while heavy telemetry is processed in the Cloud.
- Data Loop & Corner Case Handling: Cloud infrastructure continuously analyzes edge cases encountered on public roads to retrain AI models and deploy fixes back to the fleet.
- Compliance & Cybersecurity: Ensure full traceability of AI decisions and encrypted data streams to meet European AI Act requirements and ISO/SAE 21434 standards.
3. Modernize IT for Mobility as a Service (MaaS) in Tunisia and emerging markets
- AI-Driven Predictive Maintenance: Continuous telemetry analysis to detect hardware anomalies before on-road failures occur.
- Dynamic Fleet & Charging Orchestration: Algorithmic management of EV charging based on energy costs and peak urban traffic demands.
- Adaptation for Emerging Transport Networks: A practical blueprint to modernize public transport networks in Tunisia, Africa, and high-density emerging metropolises through centralized mobility data management.
Sources
