At a military installation in central California, autonomous four-seater all-terrain vehicles traverse rugged trails. This initiative is not for human operators but aims to refine AI models for navigating conflict zones.
Scout AI, a startup established in 2024 by Coby Adcock and Collin Otis, refers to itself as a "frontier lab for defense." Recently, the company announced a successful $100 million Series A funding round, led by Align Ventures and Draper Associates, following an earlier $15 million seed round in January 2025.
TechCrunch was granted exclusive access to observe Scout's training operations at a military base, the name of which remains undisclosed. The company is developing an AI model named "Fury," designed to manage military assets, initially focusing on logistical support, with plans to extend its capabilities to autonomous weaponry. According to CTO Collin Otis, this process resembles training soldiers, emphasizing the importance of a foundational intelligence level.
Scout has secured $11 million in military technology contracts from organizations such as DARPA and the Army Applications Laboratory. It is among 20 autonomy companies collaborating with the U.S. Army's 1st Cavalry Division during its training cycle at Fort Hood, Texas, with the expectation that successful technologies will be integrated into future deployments.
In its internal testing, Scout's operations team, composed of former military personnel, puts the vehicles through simulated missions in the challenging terrain of the base. Otis, motivated by his experience in autonomous trucking, recognized the need for systems capable of operating in unpredictable environments, prompting the formation of Scout.
Scout is leveraging Vision Language Action models (VLAs), an advanced technology based on large language models (LLMs) released by Google DeepMind in 2023. This innovation aims to enhance the intelligence of military autonomous vehicles. Adcock, who also serves on the board of Figure.AI, was inspired to apply this technology to military applications, collaborating with Otis to integrate cutting-edge AI solutions.
Otis notes the ease with which individuals can learn to operate drones, likening it to connecting prior knowledge with the controls. The company has been training its models using ATVs for just six weeks, transitioning from civilian vehicles to military-grade models.
Scout envisions its first major product, "Ox," as a command and control software that enables soldiers to manage multiple drones and vehicles through simple commands. This software will be tested in real-world conditions at the military base, where operators will log their experiences to refine the AI models.
The future of military logistics may see autonomous vehicles carrying supplies to remote locations, significantly reducing the burden on human personnel. As Scout continues to innovate and develop its technology, the potential for enhanced operational efficiency in military contexts is promising.
By pioneering advancements in autonomous systems, Scout AI is not only transforming military operations but also shaping the future of AI in dynamic environments.
