The South Korean government is launching a large-scale research and development program to build up the core components and software industries for self-driving vehicles. The program focuses on raising the technological competitiveness of the entire autonomous driving ecosystem — sensors, semiconductors and artificial intelligence software — rather than letting automakers lead the effort. Analysts see the move as an attempt to reduce reliance on Chinese-made parts and secure technological sovereignty in autonomous driving.
The Institute of Information & Communications Technology Planning & Evaluation (IITP), an arm of the Ministry of Science and ICT, last month commissioned a pre-planning study for a new large-scale R&D program on autonomous driving, according to information technology industry sources on the 5th. The study follows up on the Autonomous Driving AI Demonstration and Commercialization Roadmap, which the ministry approved in August at a meeting of ministers on science and technology affairs.
Through the study, IITP plans to set the strategic areas and priority tasks for the next-generation large-scale autonomous driving R&D program. Notably, the agency intends to move away from planning centered on large corporations. The project will be designed to build a full-stack autonomous driving ecosystem spanning sensors, semiconductors and AI, with parts makers, software companies and AI startups at its core.
The plan is also seen as groundwork for reducing dependence on foreign suppliers for key self-driving components. Chinese companies in particular dominate the global market for lidar, which serves as the eyes of an autonomous vehicle. Chinese firms including Hesai, RoboSense and Huawei accounted for 77% of the global lidar market as of 2024, according to overseas market research firm Yole Intelligence. Chinese lidar is also being used at domestic autonomous driving demonstration sites, partly because of its price competitiveness. Korea's self-sufficiency rate for automotive semiconductors stands at just 5%, and the government aims to raise it to 10% by 2030.
IITP said a next-generation large-scale R&D program on autonomous driving is needed to respond preemptively to the global shift toward massive AI models and software-defined vehicles, to secure national technological sovereignty and to complete a fully driverless ecosystem. "We will fully internalize core Level 4-plus technologies — autonomous driving at or beyond the fully autonomous stage — and draw up concrete milestones for expanding services nationwide," the agency said.
Industry officials expect the fundamental direction of domestic autonomous driving development to change, as end-to-end AI emerges as the core technology for realizing driverless operation. Self-driving vehicles previously moved according to rules predefined by developers, a rule-based approach. Now the mainstream approach is expected to shift to one in which AI learns the many unstructured variables of the driving environment on its own and maximizes data-driven decision-making.
