Citizen Portal
Sign In

Get Full Government Meeting Transcripts, Videos, & Alerts Forever!

Get email alerts on the Workforce Procurement Energy topic

No spam. Unsubscribe anytime.

Senators press witnesses on workforce, procurement pathways and energy limits to scale defense AI

2906915 · March 25, 2025
AI-Generated Content: All content on this page was generated by AI to highlight key points from the meeting. For complete details and context, we recommend watching the full video. so we can fix them.

Summary

Senators asked industry witnesses about workforce development, procurement entry points for industry and the energy demands of AI; witnesses recommended community training pipelines, clearer procurement pathways (citing Project Maven as a model) and rightsized models for edge deployment to reduce power needs.

Senators used the hearing to press industry witnesses on three practical constraints affecting DoD AI adoption: workforce development, procurement pathways for industry, and energy demands tied to large-scale AI systems.

On workforce, witnesses described existing and pilot efforts. Dan Tadros highlighted a St. Louis program described as training nontraditional hires to perform data development and imagery annotation and noted community-college certificate programs as a pathway into cleared AI support roles. "What we've done is taken individuals that would normally not participate in the national defense and give them an opportunity to support data development and AI development," he said.

On procurement pathways, witnesses described the process as "ad hoc" and recommended clearer, sustained mechanisms to connect industry and DoD. One witness pointed to Project Maven as an existing example that brought private-sector expertise into DoD in an organized way: "Project Maven...was to bring technology into the Department of Defense in a very aggressive manner," the witness said.

On energy and deployment, senators raised the rising compute and power needs of AI models and the need for smaller, fine-tuned models deployable at the edge. Farris described work to "build some of these models to be deployed on as small as 2 GPUs" and toward "edge deployments in laptops," reducing power and infrastructure needs. Tadros said a mix of fine-tuned smaller models interacting together can lower power requirements during fielding.

Why it matters: Congress controls training budgets and many procurement authorities, and senators used the hearing to solicit specific, actionable recommendations where legislative or appropriation levers could accelerate DoD adoption.

Ending: Witnesses urged Congress to invest in workforce pipelines, clarify procurement pathways to reduce ad hoc engagement, and encourage investment in lower-power model architectures for edge deployment.