Socar and Apex Mobility are formally entering the global autonomous-driving data market by converting 15 years of real-world car-sharing data into AI training datasets. The companies will present South Korea’s largest autonomous-driving data platform at the 2026 Autonomous Mobility Expo, a three-day event at COEX in Seoul from the 25th to the 27th, after announcing their participation on the 24th. Apex Mobility, a joint venture Socar established with Krafton in May at a total scale of 150 billion won (about $108.6 million), plans to build and supply data pipelines before expanding into robotaxi services. Socar gathers more than 100 types of vehicle data in real time from about 25,000 connected cars nationwide, covering about 1.1 million kilometers of driving each day and public-road conditions across all four seasons. Its records include more than 40,000 accident videos annually and over 220,000 specialized datasets combining accident footage with ego-motion data. Apex Mobility processes the material through de-identification, time synchronization and vision-language-model-based automated labeling to create internationally standardized training datasets. The companies are targeting edge cases as autonomous-driving development shifts from high-definition-map-based rule systems toward end-to-end AI models that process multiple sensor inputs. Apex Mobility will also unveil an in-house Ioniq 5 prototype equipped with roof-mounted LiDAR and seven multi-view cameras for producing client-specific datasets. CEO Park Jae-wook said the companies would pursue supply partnerships with domestic and international automakers and autonomous-driving companies, while combining Socar’s operational data with Krafton’s AI capabilities to build an integrated business spanning collection, processing, supply and services.