Search
Browse By Day
Browse By Time
Browse By Person
Browse By Policy Area
Browse By Session Type
Browse By Keyword
Browse Artificial Intelligence Presentations
Program Calendar
Sign In
Search Tips
Federal AI contract obligations were concentrated in a small number of states. I examine whether that geography reflects partisan alignment, a preexisting AI contractor base, or a broader capacity for federal contracting. Transaction-level records from USAspending.gov produce a fifty-state panel for FY2020-FY2024, with capacity measured from FY2015-FY2019. Fixed keyword and contextual rules applied to award descriptions identify approximately $1.65 billion in positive AI contract obligations. Obligations assignable to the states total $1.63 billion by recipient headquarters and $1.60 billion by place of performance. Virginia, California, Maryland, Massachusetts, and Texas receive about 68 percent under either definition. Poisson pseudo-maximum-likelihood estimates show positive associations with both the prior AI contractor base and the earlier non-AI federal contracting footprint. Gubernatorial alignment with the president's party has no stable association with obligations. The capacity relationships describe persistent differences between states rather than short-run changes within them. In the headquarters models, prior AI contractor capacity is positively associated with defense obligations, while the broader federal footprint is positively associated with civilian obligations. Pre-period measures of technical employment and data-processing infrastructure do not attenuate these findings. The results connect procurement capacity to the geographic distribution of federal AI obligations and suggest that efforts to broaden participation should consider access to federal procurement alongside technical development.