Nvidia expands AI ties with Hyundai, Samsung, SK, Naver (techcrunch.com)

Nvidia CEO Jensen Huang’s visit to South Korea marks a major expansion of partnerships with Hyundai, Samsung, SK, and Naver and a government-led push to scale national AI infrastructure. South Korea will acquire over 260,000 of Nvidia’s latest GPUs — roughly 50,000 earmarked for public initiatives (a national AI data center and domestic foundation-model work) and 200,000+ allocated to corporations to accelerate industry-specific models, AI-driven manufacturing, robotics, and autonomous mobility. The move follows recent U.S. technology deals with allies and underscores geopolitical and industrial momentum behind large-scale AI compute deployments. Technically, the deals bind Nvidia’s Blackwell GPUs, Omniverse simulation/digital-twin stacks, and interconnect work (NVLink Fusion / XPU plans) to concrete industrial use cases: Samsung and Nvidia will build an “AI Megafactory” using >50,000 GPUs and Omniverse for real‑time production optimization; Samsung and Nvidia are collaborating on HBM4 memory for next‑gen AI systems; Hyundai will deploy 50,000 Blackwell GPUs for training/validation in mobility and robotics; SK will create an enterprise manufacturing AI cloud; Naver is developing a “Physical AI” platform to connect digital models with real sites. Joint AI‑RAN work with telcos and ETRI aims to embed AI at the network edge for performance and power gains. For the AI/ML community, this accelerates access to massive, production‑grade compute, richer simulation datasets, and tighter hardware‑software co‑design across manufacturing, telecom, and mobility.