In September 2026, the leaders of the world’s largest artificial intelligence companies publicly called for a slower pace of development, and the American administration rejected the appeal in under a day. This analysis reconstructs the week in which the debate over AI safety became a question of global power: Beijing’s response, the data showing the gap between American and Chinese models narrowing to 2.7%, the physical infrastructure underpinning the entire competition, the case of Russia — a state that has lost the race for models yet remains militarily relevant through the drone autonomy tested in Ukraine — the accelerating integration of AI into the American military, Europe’s postponement of the AI Act’s most demanding obligations, and the next frontier, where artificial intelligence and quantum computing are already beginning to reinforce one another.