A self-driving car glides through the complex streets of Seoul. It is a marvel of engineering, a testament to Korea’s manufacturing might. However, the real revolution is not the vehicle itself. It is the invisible torrent of data the car generates, processes, and learns from every second. This distinction is at the heart of a critical debate about the future of Physical AI in Korea.

The AI paradigm has moved beyond text and images. We are now entering the era of ‘Physical AI,’ where algorithms interact directly with the physical world through robotics and autonomous systems. Kim Geon-woo, head of Kakao Mobility’s Future Platform Economy Research Institute, recently argued that this shift represents an industrial transformation larger than the mobile revolution of the last 15 years. For instance, models that process vision, language, and action in a single loop are nearing commercialization. This matters for businesses because it signals the opening of entirely new markets where digital intelligence controls physical assets. The age of purely digital AI is over.

The Data Flywheel Challenge for Physical AI in Korea

At the core of this new era is the ‘data flywheel.’ This is the virtuous cycle where a service collects data, uses it to improve, and in turn attracts more users who generate even more data. Autonomous vehicles are the ultimate example. ‘A single self-driving car generates data on a terabyte or even petabyte scale,’ explained Kim. Therefore, the ability to process and learn from this massive data flow is the key variable determining industrial competitiveness. For investors, a company’s strategy for building and controlling this data flywheel is a primary indicator of its long-term viability. Data is the new currency, and autonomous cars are its primary engine.

Korea has seen this movie before. During the smartphone boom, Korean firms sold hardware in volumes comparable to Apple. Nevertheless, Apple’s profits were over four times higher. The reason is simple: Apple controlled the software platform—the operating system and the App Store. ‘The software platform is what captures the profit,’ Kim noted. Experts now warn that the same dynamic will play out in the Physical AI space, where software and subscription services will become the main revenue drivers. This risk of value leakage to foreign platforms is already a reality in Korea’s media and e-commerce sectors. If this pattern extends to offline services managed by AI, the economic impact could be far greater. History shows that selling the device is not the same as owning the ecosystem.

Regulatory Bottlenecks and Global Competition

A major obstacle is regulation. According to industry voices, Korea’s data rules are stuck in the smartphone era and are ill-suited for the demands of Physical AI. By contrast, other countries are moving faster. The United Kingdom, for example, allows nationwide autonomous driving tests as long as basic conditions like a safety driver and insurance are met. Meanwhile, Japan is aggressively easing its rules to combat labor shortages caused by an aging population. This regulatory gap has tangible consequences. Kim pointed out that the valuation of a British autonomous driving startup, once on a similar footing with its Korean peers, has grown to be 100 times larger, a difference he attributes mainly to the regulatory environment. For international firms, this means the Korean market, despite its tech-savvy population, presents higher barriers to entry for large-scale deployment. The race is being won or lost in regulatory filings, not just in labs.

To avoid falling behind, two solutions have been put on the table: decisive deregulation and the formation of a powerful domestic industry alliance. This would involve collaboration between Korea’s giants in automaking, semiconductors, telecom, and platforms to build an integrated ecosystem. The stakes are higher than just corporate competition. As Kim emphasized, this is a matter of national infrastructure and industrial sovereignty. He called the present moment a ‘golden time’ for preemptive policy action. Without a fundamental shift toward a software-first mindset, Korea’s hardware prowess may not be enough to secure a leading role in the next technological frontier.