SK Telecom gathered the CEOs of eight Korean robotics and industrial AI startups at its Seoul headquarters, alongside executives from SK hynix, SK AX and the SK Group AI committee, to build a commercial ecosystem around physical AI. The group covers the robotics value chain, including robot training data, foundation models, humanoids, industrial arms, wearables, and factory software, with SKT providing compute infrastructure, digital twins, semiconductor relationships, and access to real industrial sites inside the SK Group. SKT’s earliest use case is robot training, planning to copy real facilities into digital twins, run thousands of simulated tasks inside its AI data centers, and deploy robots pre-tested against a customer’s actual operating environment. The first version of the platform is expected in the second half of 2026, with commercialization planned for 2027. SKT has said physical AI should contribute more than 10% of its B2B AI revenue within three years. The initiative focuses on selling the expensive pre-deployment work—data preparation, simulation, model training, validation and commissioning—so that a robot arrives ready to work, rather than selling gigawatts of compute capacity.