Although few fully autonomous systems currently operate in combat, one startup is determined to change that. Swarmbotics AI is developing coordinated robotic units—known as “swarms”—that work together to enhance situational awareness and tactical performance on the battlefield. Essentially, these swarms consist of autonomous robots that synchronize their actions to achieve collective goals.
Swarm robotics has long been relegated to the realm of science fiction, but Swarmbotics AI is now emerging from stealth mode to make that vision a reality. Their product ecosystem, dubbed ANTS (Attritable, Networked, Tactical Swarm), is designed to potentially revolutionize battlefield logistics by offering versatile, deployable platforms.
The company was founded last summer by Stephen Houghton and Drew Watson, who first met while working at Embark Trucks, an autonomous trucking firm. Houghton, a former Marine Corps officer, contributed significantly to Cruise’s growth—from a startup of around 40 employees to a giant with approximately 3,000 staff—before leading autonomous vehicle and robotics initiatives at Amazon Web Services. Watson brings a wealth of experience from his time at the CIA, followed by roles at NASA’s Jet Propulsion Laboratory where he directed software development for the Valkyrie Humanoid Robot program. After Embark was acquired by Applied Intuition, the duo saw an opportunity to leverage their expertise in a new venture.
“We became fascinated with the potential of robot swarms,” Houghton remarked. He explained that much of the private sector’s focus has been on developing expensive, high-precision autonomous systems like robotaxis—systems that need to perform flawlessly alongside humans. In contrast, Houghton and Watson envisioned applications where more affordable, less sophisticated robots could be quickly mass-produced and deployed.
Research in swarm robotics remains a relatively new field, primarily explored in academic environments and focused on UAV behavior. In real-world conflicts such as the war in Ukraine, drones are widely used, but these systems are typically human-operated and lack the autonomous cooperative capabilities envisioned by Swarmbotics.
Houghton acknowledged that both aerial and ground swarms offer significant potential, each with its own set of challenges. While aerial swarms benefit from simpler navigation, ground-based robots tend to have longer endurance and can carry heavier payloads. Swarmbotics is initially focusing on ground platforms, introducing two models: the larger, hybrid-electric, autonomous ATV named Haul ANT, and the more compact, 60-pound Fire ANT. These platforms serve as mobile power sources that can support various payloads—ranging from intelligence and surveillance to electronic warfare, kinetic operations, or medivac—aimed at maximizing flexibility and interoperability for the Department of Defense.
The advantage of using lower-cost robotic units is their rapid scalability. “With these affordable robots, you can quickly deploy them in numbers, reaching into the hundreds or even thousands,” Houghton explained. This scalability enables a diverse range of swarm behaviors that can be adapted to different tactical scenarios.
This approach aligns with a broader trend within the Department of Defense, which is increasingly favoring cost-effective, high-volume solutions through programs like Replicator. Rather than investing in the most expensive, top-performing sensors, Swarmbotics is focused on striking a balance between performance and affordability—a strategy that makes large-scale deployment feasible.
To date, Swarmbotics has secured over $4 million in funding from investors such as Quiet Capital, Silent Ventures, LMNT Ventures, and Soma Capital, with most of the capital directed toward engineering efforts. While the company is currently prioritizing the development and demonstration of advanced technology to DOD acquisition officers, they plan to eventually expand their market to include commercial sectors.
Now an 11-person team, Swarmbotics has already rolled out the initial platforms of the ANTS ecosystem. Their ongoing efforts are centered on integrating additional agents, refining swarm coordination, and enhancing the autonomy of each robot. The ultimate goal is to develop a suite of operational “plays” that users can select from to direct the swarms’ actions based on specific mission requirements.
