Google Reportedly Pitches Custom AI TPUs to South Korea in Direct Challenge to Nvidia

A September 22 report from Seoul Economic Daily indicates Google met with South Korean officials on September 9 to discuss supplying Tensor Processing Units (TPUs) for the country's sovereign-AI infrastructure. The proposed deployment targets Nvidia's massive dominance in high-performance accelerators outside hyperscaler clouds, pitching TPUs as a strategy to materially lower AI-compute costs. South Korea's interest stems from a broader push to avoid reliance on a single foreign hardware supplier and retain control over its national compute strategy. Nvidia has spent years building a structural advantage through its CUDA software platform, meaning government buyers face significant migration and maturity costs if they diversify their silicon. Market positioning data shows 275 hedge funds holding Alphabet at the end of Q2 2026, with Berkshire Hathaway increasing its Class A position by about 45% during the quarter.