Pentagon Grants Classified AI Access to Seven Tech Giants After Anthropic Break

Civil–military technology pacts have repeatedly reshaped U.S. warfare: the 1942–45 Manhattan Project fused private labs with federal command, while 2018’s Project Maven (which saw 3,000 Google employees protest drone-targeting contracts) previewed today’s AI backlash. The new multi-vendor framework signals a structural shift from bespoke defense platforms toward cloud-based, rapidly updatable commercial models—a trajectory visible since DARPA’s networked-warfare push in the 1990s and the 2004 shift to contractor-run counter-insurgency analysis in Iraq. By distributing contracts across seven firms, the Pentagon hedges against ‘vendor-lock’ and China’s state-directed AI surge, yet it also deepens the century-long trend of privatizing the means of command and surveillance, raising perennial questions of democratic oversight last debated during the 1975 Church Committee. Whether this moment is remembered as another incremental step or as the 2020s equivalent of the 1957 Sputnik shock will hinge on how autonomous these systems become and whether legal guardrails—now phrased but untested—hold over the coming decades of algorithmic warfare.