Jensen Huang’s Visit to Korea (Part 2): The Real War of AI

Let's learn more about Nvidia CEO Jensen Huang's visit to South Korea

— The Bright and Dark Sides of Downstream Industries: Automotive, Robotics, and Data Centers — As explored in Part 1, the first battlefield of the AI race was upstream (back-end industries), centered around semiconductors and supply chains. However, NVIDIA CEO Jensen Huang’s five-day visit to South Korea in early June 2026 signals another massive shift in the tech industry: AI is now stepping out of the virtual world of data centers and entering the physical world. The consecutive meetings with heads of major Korean conglomerates during his visit (the so-called "Samso Meetings," etc.) strongly resemble a strategic consensus to align directions and jointly respond to building a global AI ecosystem. As spectacularly as the technology is advancing, its actual integration into industrial sites has brought "massive real-world resistance and institutional uncertainty" to the surface. The core message piercing through this trend is clear: "AI is entering the downstream phase (front-end application industries) where it drives actual industries, and the real war begins now." Source: NVIDIA Corporation 1. Core Shift: The Emergence of "Physical AI" The most critical keyword defining this visit and recent tech trends is "Physical AI." This refers to the stage where AI perceives, makes decisions, and connects them directly to physical actions . ▶ Past AI vs. Present AI (Conceptual Comparison) Past AI: Operated primarily in digital (virtual) environments, limited to analysis and information generation such as search, recommendations, and Generative AI (text/image creation). Present AI: Expanding into the physical (real) world, executing environment assessment and directly controlling actions in robots, vehicles, and smart factories (intelligent production plants). ▶ Shift in the Value Chain Weight Upstream (Semiconductors \ Infrastructure) → Midstream (AI Models \ Software) → Downstream (Real-world Industrial Application) Moving beyond semiconductor and model development, the core of global competition has shifted to "who can seamlessly implant and apply AI into more real-world industries." 2. Automotive, Robotics, & Data Centers: The Bright and Dark Sides of Korea-Centric Downstream Transition Possessing a robust manufacturing base, advanced IT infrastructure, and cutting-edge semiconductor capabilities, South Korea is evaluated as a crucial partner for NVIDIA's downstream strategy. ① Automotive Industry: The Transition from "Vehicle → Computer → AI Platform" and "Human Resistance" Automobile manufacturing is the sector where Physical AI is being commercialized most rapidly. Structure (Cloud-to-Car Connectivity): DGX (Data Center GPU Instance): Trains massive AI models in large-scale data centers. Omniverse: Creates a Digital Twin—an exact virtual replica of the real world—to pre-verify autonomous driving simulations and safety. DRIVE AGX: An in-vehicle AI chip that assesses road conditions in real-time and controls driving operations. Collaboration with Korean Enterprises: Ongoing discussions target potential partnerships with Hyundai Motor Group and its autonomous driving joint venture, Motional , for Level 4 robotaxis and next-generation mobility. Concurrently, the role of domestic automotive electronics (infotainment/components) suppliers is expected to expand. The overarching goal is transforming traditional manufacturers into "Mobility AI Platform Companies." ※ Real-World Concerns: Vacuums in Safety and Liability: Unlike rapid technological advancements, legal and institutional standards regarding liability during accidents in Level 4+ autonomous driving remain ambiguous. It is a prime example of legislation failing to keep pace with technology. Job Displacement and Social Conflict: The full-scale introduction of robotaxis or autonomous freight trucks directly threatens the livelihoods of taxi and truck drivers. In Korea, intense opposition against Physical AI and workplace automation has already formed around labor unions (e.g., Hyundai Motor Union). This foreshadows severe labor-management conflicts and high social costs during adoption. ② Robotics Industry: "AI’s Next Battlefield" and "Labor Market Shocks" Robotics, a field heavily emphasized by CEO Jensen Huang, is the domain where AI can most intuitively prove its value. Essence of the Change: While past robots relied on "fixed-motion automation" repeating predefined trajectories, modern robotics is evolving into "AI-based adaptive systems" that perceive environmental changes and modify paths autonomously. Industrial Synergy: Comprehensive discussions are underway with Korean corporations possessing robust hardware manufacturing capabilities (such as Doosan and LG ) to link ecosystems across AI-powered factory automation, logistics, and service robotics. ※ Real-World Concerns: Deploying Physical AI into manufacturing and logistics lines maximizes labor productivity but triggers a sharp decline in production-line jobs. In Korea's employment structure, this easily morphs into the frame of "AI adoption = restructuring," acting as the largest bottleneck restricting the speed of domestic implementation. ③ Data Centers: "The Foundation of All Industries" and "Data Sovereignty Risks" For vehicles and robots to move, a massive computing engine—the AI Data Center—is essential behind the scenes. Structural Transformation: Past data centers were simple "Cloud Servers" that stored and distributed data. Today, they are transforming into "AI Factories" that take raw materials (data) and manufacture value-added outputs (intelligence) . Korea’s Role: Amid a nationwide push for "Sovereign AI" (maintaining independent national AI capabilities) , Naver , alongside domestic telecom and IT service (SI) companies, is striving to expand a localized ecosystem utilizing NVIDIA's infrastructure. ※ Real-World Concerns: Deeper infrastructure collaboration with global Big Tech platforms fuels security and legal anxieties regarding the potential leakage or dependency of core domestic data and industrial assets overseas. Compared to the speed of global tech standardization, country-specific data security and regulatory laws are not yet clearly established, leaving companies exposed to highly volatile regulatory risks. 3. Industrial Convergence: Connected into a Single AI Ecosystem Automotive, robotics, and data centers are not isolated industries; they are bound together in a massive "AI Feedback Loop." Domain Role Description Data Center (Infrastructure) Learning Trains large-scale models based on data collected from reality. Automotive (Mobility) Mobile AI Drives on real roads, collecting real-time environmental data. Robotics (Mfg / Logistics) Action-oriented AI Executes physical tasks in factories or daily life, accumulating operational data. Integration Mechanism: [Collect Real-world Data] → [Train in Data Centers] → [Optimize \ Execute on Devices] → [Regenerate New Data] This loop repeats infinitely, exponentially advancing AI performance. The Paradox of Job Reduction: The issue lies in the fact that as this performance optimization loop accelerates, the "human job reduction loop" accelerates alongside it. Because every nation faces the political burdens of tax revenue collection and employment stability, we face a dual confrontation: technology advancing at the speed of light on one side, while actual implementation and new industrial regulations sit in a tense, ambiguous stalemate on the other. 4. India as a Detour: A "Breakthrough" or "Another Uncertainty"? At this juncture, India must not be interpreted merely as a "consumer market to sell products." This aligns with the recent South Korea-India advanced digital cooperation framework (Digital Bridge Framework). India’s Core Role: India possesses the world’s largest software talent pool and serves as a core axis in the "Post-China" supply chain diversification strategy. The core question is not "Should we sell finished goods to the Indian market?" but rather "How do we link India's workforce and infrastructure within our AI and manufacturing supply chains to diversify risk?" However, India should not be viewed blindly as a "regulatory sanctuary" or a flawless breakthrough. The internal Indian market harbors its own massive structural uncertainties. The Boomerang of Youth Unemployment: India faces severe youth unemployment relative to its massive population scale. If AI and physical automation rapidly cannibalize typical Indian white-collar (IT outsourcing) and blue-collar jobs, the Indian central government is highly likely to brandish strict regulations to protect domestic employment. Complex Legal and Regulatory Landscape: India presents wide regulatory disparities between the central government and its 28 state governments. Particularly, as the detailed enforcement rules of India’s Digital Personal Data Protection (DPDP) Act materialize, interpretation and regulatory stringency vary heavily by state, forcing corporations to navigate multi-layered legal risks. Infrastructure Limitations (Test-bed vs. Large-scale Hubs): Moving past the "test-bed" phase of piloting technologies near major metropolitan areas to constructing large-scale AI physical manufacturing and computing hubs with guaranteed operational stability is an entirely different challenge. Given chronic power grid instabilities and logistics infrastructure constraints, doubts remain over whether large-scale hubs can be established in a timely manner. 5. Practical Insights: Challenges and Choices for Korean Enterprises ▶ Opportunities Participation in constructing global AI factory (data center) infrastructure. Early implantation of AI into Korea's stronghold domain: hardware manufacturing (automobiles, robots). ▶ Strategic Directions Diversifying Downstream Domains: Expand AI application scopes rapidly centered on mobility, smart logistics, and factory automation, moving beyond simple home appliance manufacturing. Domestic (Tech Consolidation & Social Compromise): Keep core technologies and high-value manufacturing capabilities domestic, but prioritize establishing a "social protocol" between labor, management, and government to cushion the employment shocks of AI adoption. India & Global (Risk Diversification Bases): Utilize India as a base to secure software engineer pools and hedge global regulatory risks , but design a "localized, close-knit connectivity strategy" that precisely factors in India’s state-level regulatory variances and infrastructure limitations. Securing Independent Capabilities: Actively leverage global Big Tech AI platforms (hardware and software standards), but internalize software optimization capabilities to avoid becoming entirely dependent on a specific platform. ▶ Potential Risks There is a distinct danger of becoming trapped in a sub-contracting structure—surrendering core high-margin revenues to global platforms while remaining a low-margin hardware assembler or component supplier. Therefore, developing proprietary service ecosystems must be explored in parallel. Closing Thoughts If Part 1 was a war over the "engines and components" (Upstream) to create AI, the core of Part 2 is a war of friction—where AI equipped with those engines steps into the physical world and collides with the massive walls of human workers, legacy laws, and country-specific regulations (Downstream). For now, your specific business might seem detached from this trend. The current speed of technological evolution is dizzyingly fast, and experts remain divided on exact commercialization timelines. However, one stark fact remains undeniable: no one disputes the massive, overarching trajectory of "Physical AI and robotics diffusing across all industries." Behind the glamorous moves of global tech leaders lies a calculated, cutthroat race to dominate this macro-ecosystem. Individual companies cannot alter the direction of this monumental epochal shift. However, now is the time for strategic flexibility—to accurately read which way the wind of change is blowing and smartly harness its momentum.