A senior U.S. national-security voice is warning that shortages in AI compute and data infrastructure would be “catastrophic,” linking the Pentagon’s future targeting and intelligence workflows to data-center capacity. Retired Air Force Lt. Gen. David Deptula, now dean of the Mitchell Institute for Aerospace Studies, argues in a Washington Post op-ed that data storage, movement, and security underpin nearly every defense function. Deptula also points to disputes and dependencies across the defense AI ecosystem, citing the Pentagon’s reliance on AI platforms and the sensitivity of intelligence pipelines if key software systems fail or restrict outputs. The op-ed frames data infrastructure as a strategic asset rather than an IT cost center. For universities and research organizations, the development adds urgency to campus-level data governance, GPU/compute planning, and cybersecurity hardening—especially where institutions partner with defense-related contractors or run AI research pipelines that require sustained compute. The piece also notes public backlash to new AI data centers, raising the likelihood that permitting, power supply, and community opposition could become limiting factors for future AI research capacity.