RAND Corporation researchers argue that the United States must not commit prematurely to any single artificial superintelligence strategy. Instead, America should pursue a freedom of action approach that invests immediately in verification, monitoring, and resilience capabilities to keep strategic options open.