Colby Adcock’s Scout AI raises $100 million to train models for war. we visited the boot camp

At a U.S. military base in central California, a four-seater all-terrain vehicle roams a hillside path. This is training, but not for the people in the vehicle. An effort to train AI models to enter conflict zones.

The autonomous military ATV is operated by Scout AI, a startup founded in 2024 by Colby Adcock and Colin Ortiz that calls itself “Frontier Lab for Defense.” The company announced Wednesday that it has raised $100 million in a Series A round led by Align Ventures and Draper Associates, following a $15 million seed round in January 2025.

Scout invited TechCrunch to take an exclusive tour of the training operation at the military base, which he asked not to be named.

The company is building an AI model called “Fury” to operate and command military assets. Initially it will be used for logistics support, but soon it will also be used for autonomous weapons. CTO Collin Otis likens the effort, which builds on existing LLMs, to training soldiers.

“[Soldiers] “They start at age 18, and in some cases they start after they graduate from college. So we want to start with that base level of intelligence. It’s helpful to start with people who are already invested and think, ‘Hey, what do we have to do to teach this as an incredible military AGI rather than just a broad intelligence AGI?'” Ortiz told TechCrunch.

Scout has won military technology development contracts totaling $11 million from organizations such as DARPA, the Army Applied Research Laboratory, and other Department of Defense customers. The company is one of 20 autonomous companies whose technology is being used during routine training by the U.S. Army’s 1st Cavalry Division at Fort Hood, Texas, and is expected to bring a proven product during the unit’s next deployment in 2027.

Internal testing of the Scout sees the rubber make contact with dirt on the base’s hilly terrain, while the company’s operations team, led by former soldiers, tests the vehicle in simulated missions.

Self-driving cars are starting to be seen in more cities around the world, but self-driving cars operate in a more structured environment with rules. Operating autonomously on unmarked trails and off-road is another challenge entirely. Ortiz, who previously worked at the self-driving truck company Kodiak, said he was inspired to start Scout after realizing that the systems he helped develop were not intelligent enough to operate in unpredictable combat zones.

A new approach to autonomy

Scout focuses on a newer autonomous technology called Vision Language Action Model (VLA), which is used to control robots based on LLM. The technology, first released by Google DeepMind in 2023, has seeded robotics startups such as Physical Intelligence and Figure.AI, a humanoid robot company led by Adcock’s brother Brett.

Colby Adcock, a Figure executive, said his experience convinced him of the opportunity to bring broad intelligence to the military’s growing fleet of self-driving vehicles. His brother introduced him to Otis, who was advising Figure, and they set out to apply modern AI to military solutions.

“Right now, I hand you a drone controller, put on a headset, and you can learn to fly a drone in minutes,” Ortiz said. “It’s really just learning how to connect prior knowledge to some of these little joysticks. It’s not a huge leap. This is how we think about VLA and why VLA is unlocked.”

In fact, I had the opportunity to drive a Scout ATV over rutted trails, and the terrain was challenging, with steep hills, loose sand around turns, disappeared railroad tracks, and confusing intersections. I’m not an experienced ATV driver, but I did pretty well on my first try (if I do say so myself). That’s the kind of general intelligence the company looks for in its models, and the company has been training with these ATVs for just six weeks. At first we used civilian ATVs.

I also rode an autonomously controlled ATV and felt the difference in accelerating faster than a human with passenger comfort in mind. The management team, like the training drivers, noted how the vehicle moved to the right on wide roads, but stayed in the center on narrow roads. Also, when they became confused, they would suddenly slow down to think about their next move. That happened several times as the ATV carried us on a 4-mile loop back to base.

Although VLA is new enough that no company has deployed it in an operational environment yet, “the technology is good enough to be tested in the field with soldiers to find the most effective method for the U.S. military,” said Stuart Young, a former DARPA program manager who worked on ground vehicle autonomy. Like other autonomous companies, Scout’s full stack also includes deterministic systems and other types of AI to enrich the agent’s capabilities.

Young left DARPA to join Field this month after managing a program called RACER that asked companies to develop high-speed, self-driving off-road vehicles, and helped seed the field in the same way the organization’s Grand Challenges boosted self-driving cars. Two competitors in the space, Field AI and Overland AI, spun off from that program, and Scout later joined.

The first application of ground autonomy will be autonomous replenishment, Scout executives and military engineers say. Transporting water and ammunition to distant observation posts, or in convoys of six to 10 self-driving cars followed by a manned truck, can save valuable human labor for more important missions.

Brian Maswich, an active-duty infantry officer serving as a Scout military fellow, recalled a recent exercise in Alaska where he led a supply convoy in pitch darkness, hoping self-driving cars would help him.

Adds intelligence to the Army’s motor pool

Scout sees itself as a software company primarily building the intelligence layer of military machines. We’re not going to build self-driving cars ourselves, we’re going to build on top of them.

Adcock expects the startup’s first product to be widely adopted will be something called Ox, a command-and-control software bundled with enhanced computer hardware such as GPUs, communications and cameras. This is intended to allow individual soldiers to coordinate multiple drones and autonomous ground vehicles using prompts such as “Go to this waypoint and observe enemy forces.”

However, in order to make the software work, training on actual vehicles is required, so a foundry, a training range, was set up on a military base. There, drivers spend eight-hour shifts putting the ATV through its paces, then using a reinforcement learning system to record where they had to take over, which is used to improve the model. Base commanders have even asked the company’s ATVs to replace security patrols.

One of the hypotheses that Scout is testing is that VLA enables this relatively limited data set and simulation training data to provide a fully functional driving agent. For example, this vehicle looks comfortable on the trail, but is not ready to operate completely off-road.

Scout also trains with drones for reconnaissance and defense, giving them intelligence through vision language models.

The startup is working on a system where groups of military drones fly on a larger “quarterback” platform that provides more computing resources to command them. For example, a drone could search for hidden enemy tanks in a geographic area and possibly attack them without human intervention. Ortiz argues that an alternative approach in such a scenario could be indirect artillery fire, but this would be inaccurate compared to drone attacks.

Autonomous weapons are a hot topic in defense technology policy, but experts say the concept is old, with heat-seeking missiles and landmines having been used in warfare for decades. The question for engineers is how the weapon will be controlled, said Jay Adams, a retired U.S. Army captain who heads the Scout operations team.

Adams points out that the company’s military drones can be programmed to only attack threats in specific geographic areas, or only after human confirmation. He also said it was unlikely that an autonomous weapons platform would open fire in fear like the 18-year-old soldier did.

VLA is also expected to improve targeting. Scout says its models are pre-trained on specific military data, in case it collides with an enemy tank during a resupply mission, for example. Lt. Col. Nick Rinaldi, who oversees scouting research at the Army Applied Research Laboratory, said automated targeting is difficult and unlikely to be used outside of a constrained environment in the short term, but it is a promising technique for investigation because of the potential VLA can provide to infer about threats.

Adams says the ability of drones to identify their own targets is key to future warfare. Although Russia’s invasion of Ukraine has sparked strong interest in drone warfare, he believes that human control of individual UAVs cannot scale sufficiently for the United States to counter large numbers of low-cost unmanned systems in the event of a threat from the U.S. military.

Mission to counter anti-military atmosphere

Like many defense startups, Scout has a mission in mind, and its executives won’t hesitate to criticize companies that are reluctant to hand over technology to the government. For example, Google reportedly pulled out of a Pentagon competition to develop a control system for autonomous drone swarms, a feature Scout is also working on.

“AI people don’t want to work with the military,” Ortiz told TechCrunch, referring to Anthropic’s dispute with the Department of Defense over terms of use. “Neither of them accept running agents on one-way attack drones or running agents on missile systems.”

Nevertheless, Scout is using an existing LLM as the basis for building its agents, but did not reveal which LLM it is. Otis said it has contracted with “a very well-known hyperscaler” to provide pre-trained intelligence to Scout’s underlying model. He also declined to say whether the company uses an open weight model like those offered by Chinese companies. Many companies that rely on AI inference are built on openweight models because they are cheaper compared to what frontier labs like Anthropic and OpenAI offer.

Scout plans to address this problem by building its own models from scratch over the next few years, and the founders say much of the capital will go toward their training and computational costs. In fact, Ortiz wonders if Scout will be able to defeat existing leaders with AGI, since Scout’s models constantly interact with the real world.

“There is an argument in the AGI community that you can only increase your intelligence by reading the Internet, and that most intelligence comes from interacting with the world,” Ortiz said.

So does that mean Adcock is competing with his brother’s humanoid robot army in figures? No, says Ortiz, but “we can scale faster because our customers have the assets,” referring to the Department of Defense.