Browse
Log in / Register
Share
Automakers are racing to turn massive vehicle data into a competitive advantage, and Hyundai Motor Group is betting big on a “data flywheel” strategy. On August 26, 2025, during its 2025 CEO Investor Day, the South Korean giant outlined a plan to continuously collect and analyze data from its global fleet, enabling over-the-air updates that improve autonomous driving, infotainment, and overall software-defined vehicle (SDV) performance.
The approach is a textbook virtuous loop: each vehicle becomes a sensor, feeding real-world data into machine learning models. Those models then refine features, which are pushed back to cars via OTA updates, making the fleet smarter over time. Hyundai’s vision is to eventually standardize sensor architecture around the Nvidia ecosystem, ensuring seamless data usage across its main brand, Kia, its software subsidiary 42dot, and autonomous driving unit Motional.
The flywheel starts with Hyundai’s flagship Grandeur sedan in South Korea, equipped with the Pleos Connect infotainment system and Gleo AI, a generative AI assistant. Data from these vehicles will first enhance Gleo AI’s capabilities—centered on vehicle controls and convenience—and later link with other application services. This is a departure from the typical automaker approach, where data is often siloed. Hyundai instead aims to create a unified stream that feeds both infotainment and autonomy.
On the autonomous driving front, Hyundai plans a staged rollout. By 2028, it will introduce Level 2+ autonomous driving on its first mass-produced SDV, leveraging its ongoing collaboration with Nvidia. The accumulated data from these vehicles will then advance Atria AI, an end-to-end autonomous driving system developed by 42dot. Atria AI is already being deployed in South Korea by the end of 2025 to collect “edge cases”—unexpected and unusual road situations—which are critical for training robust AI models. Hyundai expects to eventually deploy Atria AI in millions of vehicles to support both Level 2+ and Level 4 autonomy.
To handle the massive data volume, Hyundai is building its own large-scale data center. The Saemangeum AI Data Center in South Korea, a 100-megawatt facility capable of housing over 50,000 GPUs, is expected to go online in stages starting in 2029. This will not only process autonomous driving data but also connect data from Hyundai’s global production system and its integrated in-house AI platform. The company believes this infrastructure will give it a significant edge in SDV development and autonomous driving performance.
Value-Added Perspective: Hyundai’s data flywheel mirrors the strategies of Tesla and Waymo, but with a distinct focus on hardware standardization and OEM-to-supplier integration. Unlike Tesla’s closed-loop within its own fleet, Hyundai is leveraging multiple brands and subsidiaries to generate a wider variety of driving scenarios. Industry analysts at Gartner have noted that such a strategy could accelerate AI learning if the data is properly labeled and curated—a challenge Hyundai is addressing by putting edge-case collection at the forefront. Additionally, the collaboration with Nvidia ensures that Hyundai’s sensor architecture aligns with the most widely adopted AI computing platform in autonomous driving, reducing fragmentation.
The road ahead is ambitious. Beginning in 2029, Hyundai expects the volume of autonomous driving data from its connected fleet to grow exponentially. With the Saemangeum data center coming online, the company is preparing for a future where data is the fuel for vehicle intelligence. For now, the flywheel is just starting to spin, but Hyundai’s bet is clear: those who master data collection and processing will lead the SDV era.
