Scout AI raised a $100M Series A led by Align Ventures and Draper Associates to train an AI model that drives military vehicles and commands drones, the company said Wednesday. The round follows a $15M seed in January 2025 and lands alongside $11M in development contracts already booked with DARPA, the Army Applications Laboratory and other Department of Defense customers. Founded in 2024 by Coby Adcock and Collin Otis, Scout AI calls itself a frontier lab for defense and is building toward autonomous weapons after starting with logistics.
The flagship model, called Fury, is being trained on four-seater all-terrain vehicles at an unnamed U.S. military base in central California. Scout is also one of 20 autonomy companies whose technology is being used by the U.S. Army's 1st Cavalry Division at Fort Hood in Texas, with the unit expected to carry forward systems that prove themselves into its 2027 deployment. The training site, which Scout calls Foundry, runs drivers through 8-hour shifts on a 6.5 km loop.
Fury is built on Vision Language Action models, an architecture first released by Google DeepMind in 2023 that has since seeded robotics startups including Physical Intelligence and Figure.AI. Coby Adcock sits on the board of Figure, the humanoid robot company run by his brother Brett Adcock, and that exposure pushed him toward applying VLAs to ground vehicles. Otis, previously at autonomous trucking company Kodiak, said he started Scout after concluding that structured-road autonomy stacks weren't intelligent enough for unpredictable terrain.
Key facts
- 01Scout AI raised a $100M Series A led by Align Ventures and Draper Associates, following a $15M seed in January 2025.
- 02The startup holds $11M in development contracts with DARPA, the Army Applications Laboratory and other DoD customers.
- 03Scout is one of 20 autonomy companies whose tech is being tested by the U.S. Army's 1st Cavalry Division at Fort Hood, with deployment expected in 2027.
- 04Its Fury model has trained on real ATVs for just six weeks, with drivers running 8-hour shifts on a 6.5 km loop.
- 05Vision Language Action models, the architecture underpinning Fury, were first released by Google DeepMind in 2023.
Otis frames the training process as analogous to onboarding a soldier. "Soldiers start when they're 18 years old, and sometimes they even start after college, so you want to start with that base level of intelligence," he told TechCrunch. "It's useful to start with someone who's already made an investment and then say, 'Hey, what do I have to do to teach this thing to be an incredible military AGI, versus just being a broadly intelligent AGI?'"
“Scout AI's models have trained on real ATVs for just six weeks, but the company has already secured $11M in contracts from DARPA, the Army Applications Laboratory and other DoD customers.”— Jaeden Schafer
The pitch for VLAs is data efficiency. "If I handed you the controller of a drone right now and I strapped a headset on you, you could learn to fly that thing in minutes," Otis said. "You're actually just learning how to connect your prior knowledge to these couple little joysticks. It's not a big leap. That's the way to think about VLAs and why they're such an unlock." The Fury model has been training on the military ATVs for only six weeks, after starting on civilian vehicles.
Scout's first commercial product is expected to be Ox, a command-and-control software package bundled with hardened compute, communications and cameras. Ox is designed to let a single soldier orchestrate multiple drones and autonomous ground vehicles through natural-language prompts. The early use cases are unglamorous: automated resupply runs carrying water and ammunition, or convoys where a crewed truck leads six to ten autonomous vehicles behind it.
Stuart Young, a former DARPA program manager who ran the RACER off-road autonomy program and left the agency this month, said VLAs aren't yet operationally deployed by any company but "the technology is good enough to be doing that experimentation in the field with soldiers to figure out how to most be effective to U.S. forces." Two RACER alumni, Field AI and Overland AI, are now competing with Scout, which also participated in the program.
Scout positions itself as a software layer rather than a vehicle maker, betting that the intelligence stack is the durable moat as the Army's motor pool fills with autonomous platforms from multiple vendors. The company is also experimenting with munition drones flown alongside a larger "quarterback" platform that provides extra compute, a configuration that could in theory hunt enemy armor with limited human input. Targeting can be geofenced or gated on human confirmation, according to Jay Adams, a retired Army Captain who leads Scout's operations.
The autonomous-weapons question is where the politics get sharper. Lt. Col. Nick Rinaldi, who supervises Scout's work for the Army Applications Laboratory, said "while automated targeting is hard and unlikely to be used outside of constrained environments in the near term, the potential of VLAs to reason about threats make them a promising technology to investigate." That is a careful endorsement: useful, worth studying, not ready to fire on its own. VLAs have never been deployed operationally, and Scout's six-week training window on real vehicles is a thin base for systems that will eventually be asked to make life-or-death calls under degraded comms.
Scout's $100M round is a bet that the defense autonomy stack will look more like a foundation-model business than a hardware one. The economics rhyme with the broader AI capex story: companies that own the model layer capture the margin while the vehicles, drones and trucks become commoditized chassis. If Fury holds up at Fort Hood and rides into the 1st Cavalry Division's 2027 deployment, Scout will have something most defense startups still lack — a model trained on the actual job, not a simulator. If it doesn't, the $100M will have funded one of the more expensive ATV fleets in central California.
Working on something we should cover, or seeing a story we missed? Send leads, documents, or feedback to hello@aichatdaily.com. For sensitive tips, see our secure tips page for Signal and PGP options.
Spotted an error? Email hello@aichatdaily.com with the URL and the issue, or read our full corrections policy.
