Caitlin Kalinowski resigned from OpenAI, citing principle‑based objections to permissive Pentagon uses including surveillance and lethal autonomy. Her exit follows public disquiet inside tech firms over defense partnerships and came as the Defense Department confronted a sudden dependence on specific AI vendors. A senior Pentagon technology official described a 'whoa moment' when officials realized the department’s reliance on Anthropic could leave military operations vulnerable if access were restricted. That realization helped trigger rapid procurement moves and a broader scramble to diversify AI suppliers for classified and mission‑critical settings. For research offices and tech transfer leaders on campus, the episode raises questions about university partnerships with defense contractors, export controls, and ethical review processes for dual‑use AI research.