A Defense AI Startup Moves From Pitch to Field Training
Scout AI says it has raised a $100 million Series A round as it develops AI systems intended to operate and command military assets. The company, founded in 2024 by Coby Adcock and Collin Otis, describes itself as a frontier defense AI lab. What distinguishes its latest moment is not only the size of the funding round, but the way the company is training its models: on autonomous military vehicles moving through terrain at a U.S. base in central California.
The startup’s model, called Fury, is being built first for logistical support and later for autonomous weapons applications. That trajectory captures the larger pattern emerging across defense technology. Many companies begin by framing autonomy as support infrastructure, then position it as a foundation for more direct combat use once the system proves itself in operationally relevant conditions.
Training AI for War, Not General Use
Scout’s leadership describes its technical approach as building on existing large language models and then specializing them for military tasks. Otis compared the process to training soldiers: start with a base level of intelligence and then teach the system to function as what he called an “incredible military AGI” rather than a broadly intelligent general-purpose system.
That framing matters because it highlights a growing split inside AI development. Consumer and enterprise AI emphasize general capability and broad software usefulness. Defense-focused AI increasingly aims for narrower, mission-conditioned behavior under real operational constraints. In Scout’s case, that means running autonomous all-terrain vehicles through simulated missions rather than relying only on digital training environments.
