South Korean tech giants Samsung Electronics and SK Hynix have announced a massive 800 trillion won investment—roughly 518 billion dollars—to construct a new semiconductor hub in the country's southwest. This project aims to meet the accelerating global demand for AI infrastructure, including data centers and autonomous vehicles. The episode also analyzes the record-breaking "model glut" of June 2026, which saw 31 model launches, most notably OpenAI’s GPT-5.6 series—comprising Sol, Terra, and Luna—and Anthropic’s Claude Fable 5. Nina Park and Thatcher Collins explore why these updates, while impressive in benchmark testing, signal a shift toward specialized access and improved business utility through long-context handling rather than just raw reasoning power. We also discuss the intervention of the U.S. government regarding GPT-5.6’s capabilities in biology and cybersecurity, leading to a staggered release strategy. The discussion concludes by evaluating how increased competition among AI labs is driving down costs for enterprise users, regardless of whether they adopt the most recent frontier models.