Nvidia Eyes Korea's Robotaxi Pilot to Spread Self-Driving Platform

Nvidia's proposal centers on integrating its Hyperion platform and Halos safety technology into South Korea's planned Level 4 driverless robotaxi services, which MOLIT intends to launch from 2028. Hyperion combines an in-vehicle AI computer, sensors like cameras and radar, and associated software. Halos is designed to manage system errors, hazardous situations, and validate self-driving safety. If adopted in the initial Jeonnam-Gwangju pilot, this technology could be applied to Hyundai Ioniq 5-based autonomous vehicles and combined with in-house software from participants like RideFlux and Autonomous A2Z.

This initiative aligns with Nvidia's broader strategy to position its technology as a common platform for the global self-driving market. The company has expanded its offerings since entering the autonomous driving business in 2015 with Drive PX, now including AI models such as Alpamayo and simulation tools. Recent efforts include plans with Uber to expand driverless robotaxi services to 28 cities worldwide by 2028, beginning next year in Los Angeles and San Francisco.

South Korea appears to be a key market for Nvidia due to the government's plans for a significant expansion of its pilot programs. Following Jeonnam-Gwangju, MOLIT intends to select four to five additional cities in 2027 and increase the number of deployed vehicles from 200 to 3,000. This offers Nvidia an opportunity to validate its AI models and in-vehicle computing platform on real roads across multiple Korean urban environments.

However, industry sources indicate that the benefits of Nvidia's platform for Korean participants require closer scrutiny. This suggests that the value proposition of adopting Nvidia's technology as a standard is not yet universally accepted or fully understood among potential local partners.