NSA spends billions testing frontier AI models
According to sources familiar with classified intelligence estimates, the National Security Agency informed lawmakers that it is spending billions of dollars in taxpayer funds this year to evaluate and stress-test advanced artificial intelligence models for national security vulnerabilities. The classified price tag—funded through defense intelligence budgets and far outstripping prior congressional estimates of tens of millions for federal AI oversight—indicates that comprehensive government evaluation of cutting-edge models could cost tens of billions annually.
The exorbitant cost of evaluating frontier models exposes the fallacy of low-budget AI oversight: independent government testing is economically unsustainable without imposing direct regulatory fees on frontier model developers.
- –Compute costs dwarf policy budgets: While lawmakers drafted bills allocating tens of millions of dollars for oversight, running evaluations on frontier architectures in-house requires billions in specialized compute and hardware allocations alone.
- –The public-sector talent deficit: Despite housing premier government cryptographers and engineers, federal agencies cannot match the nine-figure compensation packages offered by frontier labs, exacerbating evaluation bottlenecks.
- –Industry self-policing vs. taxpayer burden: As tech giants propose private safety standards bodies to avoid government intervention, the multi-billion-dollar price tag strengthens the case for levying mandatory audit fees on frontier AI labs rather than saddling taxpayers.
DISCOVERED
1h ago
2026-09-25
PUBLISHED
2h ago
2026-09-25
RELEVANCE
AUTHOR
rdmuser