Apex Mobility will first build autonomous driving data pipelines and then expand into robotaxi services, pursuing vertical integration by combining Socar’s car-sharing operations with Krafton’s AI capabilities.

The roadmap converts Socar’s 15 years of car-sharing data into AI-ready assets, creating a high-quality data platform to support autonomous driving services.

The plan evolves from data pipelines to roboticaxi rides, targeting Level 4 autonomous ride-hailing, while Socar advances to Level 2+ autonomy in car-sharing.

Jang Hyuk, Apex Mobility’s Chief Strategy Officer, says the golden window for Korea’s autonomous driving market could be up to three years, noting progress in the U.S., China, London, and Tokyo.

Apex Mobility prioritizes edge-case data, with Socar collecting over 40,000 accident videos annually and ego-motion data at impact, generating more than 220,000 specialized datasets refined into standardized training data via automated labeling.

Industry observers expect data competitiveness to be a decisive differentiator as end-to-end approaches gain prominence, with Socar arguing its car-sharing data is more diverse than typical consumer-vehicle data.

Socar collects real-time data from about 25,000 connected cars nationwide, including dashcam footage, GPS, and steering data, totaling roughly 1.1 million kilometers driven daily across diverse road conditions.

Apex Mobility will showcase a purpose-built Ioniq 5 autonomous data-collection prototype with full sensor suites to generate customized datasets for client-specific regions and scenarios.

The exhibition featured a lidar- and seven-camera-equipped Ioniq 5 prototype to demonstrate data capabilities for AI training.

Park Jae-wook, CEO of Socar and Apex Mobility, emphasizes that real-time data quantity and quality will determine competitiveness in an end-to-end autonomous era, with plans for partnerships with automakers and robotics firms using the live data pipeline.

The strategy shifts from rules-based systems relying on precise maps to end-to-end AI models that ingest video and lidar data, using edge-case data on accidents as a critical barrier to progress.

Socar’s strategy leverages its large, diverse fleet and real-time data flow to differentiate in the autonomous driving data market by capturing a wide range of drivers and environments.