How South Korea plans to best OpenAI, Google, others with homegrown AI (techcrunch.com)

South Korea has launched a ₩530 billion (~$390M) sovereign AI initiative funding five domestic teams to build large foundational models, with progress reviews every six months that will whittle the field down to two national leaders. The aim is to reduce dependency on foreign providers, strengthen data control and national security, and commercialize models tuned to Korean language, culture and industry needs — a strategic push to compete with OpenAI, Google and others on home turf. The cohort mixes big-tech scale and startup agility with distinct technical approaches. LG AI Research’s Exaone 4.0 (32B) emphasizes hybrid reasoning, industry-specific data and chip efficiency rather than raw parameter scale. SK Telecom’s A.X 4.0 — with A. personal-agent services already at ~10M users — claims ~33% greater efficiency on Korean inputs versus GPT-4o and has open‑sourced A.X 3.1 while building GPUaaS and hyperscale data centers with AWS. Naver Cloud leverages a true “AI full stack” with HyperCLOVA X and multimodal HyperCLOVE X Think, embedding models across search, shopping and services. Startup Upstage’s Solar Pro 2 (31B) is recognized as a frontier model on Korean benchmarks and focuses on cost-effective, industry-specialized models for finance, law and medicine. Together these efforts prioritize language specialization, efficient inference, real-world data feedback loops and domain fine-tuning — a pragmatic strategy to outmaneuver global giants on relevance and operational cost rather than sheer scale.