Overview of the Partnership
Anduril Industries, a prominent defense-tech company, has revealed fresh insights into its collaborative effort with Meta to develop augmented-reality (AR) headsets tailored for military use. The project aims to reshape battlefield awareness by integrating drones, soldiers, and AI into a single, seamless network. Quay Barnett, a former U.S. Army Special Operations commander now serving as Anduril's vice president leading this initiative, describes the overarching goal as optimizing "the human as a weapons system." This cyborg-inspired vision envisions soldiers and drones sharing visual data, communicating effortlessly, and making split-second decisions as a unified force.
Two Prototypes: SBMC and EagleEye
Anduril is currently pursuing two distinct yet related AR headset projects. Each targets different aspects of future warfare, but both leverage the core technologies of augmented reality, voice control, and artificial intelligence.
SBMC: The Army's Requested System
The first project, known as the Soldier Born Mission Command (SBMC), was initiated through a $159 million prototyping contract awarded to Anduril last year. Under this agreement, Anduril is collaborating with Meta to create AR glasses that can be attached to existing military helmets. The SBMC program is designed to provide soldiers with enhanced situational awareness by overlaying critical information directly onto their field of view.
EagleEye: Anduril's Independent Venture
Alongside the Army-funded SBMC, Anduril has also launched a self-financed side project called EagleEye, announced in October. Unlike SBMC, EagleEye is a complete helmet and headset combination that the military has not specifically requested. Nevertheless, Anduril believes the military will eventually prefer this integrated design and purchase it. Both systems remain years away from deployment; the Army is not expected to choose a production model for SBMC until 2028 at the earliest, and there is no guarantee it will proceed—Microsoft previously held the lead on a similar $22 billion contract that was canceled due to viability issues.
The Core Technology: Overlay, Voice, and AI
The AR glasses for both prototypes function by projecting relevant data onto a soldier's visor. The information can range from simple cues like a compass heading to complex overlays such as real-time maps, drone locations, and AI-generated target recognition—for instance, identifying a specific vehicle. Interaction is primarily hands-free: soldiers speak to the system in natural language, issuing commands such as requesting an evacuation for an injured comrade or plotting a route that avoids restricted zones.
The Role of Large Language Models
To interpret these spoken instructions, Anduril is testing several large language models (LLMs), including Google's Gemini, Meta's Llama, and even Anthropic's Claude—despite ongoing tensions between Anthropic and the Pentagon. These LLMs translate a soldier's plain speech into executable software commands, enabling a more intuitive user experience than traditional menu-driven interfaces.
Lattice: The Central Nervous System
The underlying engine for both systems is Anduril's Lattice software platform, which aggregates data from diverse military hardware into a single unified picture. Lattice acts as the central nervous system of the operation, processing information from drones, sensors, and command centers. The Army has already committed $20 billion to integrate Lattice across its entire infrastructure, signaling strong institutional support for this approach.
Envisioning Multi-Step Operations
Barnett's team is designing the headset to handle multi-step tasks autonomously. For example, a soldier could instruct a drone to survey an area and return only after locating an artillery unit. The system would then analyze the situation and recommend a course of action—such as dispatching a nearby drone to strike—though final approval would still require human command authorization. This demonstrates a shift toward human-machine teaming, where the AI handles routine coordination while humans retain decision-making authority over lethal actions.
Timeline and Challenges
Both SBMC and EagleEye are still in early prototyping phases. The Army's decision on whether to move SBMC into full production is not expected until 2028, and past experience with similar programs (like Microsoft's canceled contract) underscores the technical and financial risks. Despite these uncertainties, Anduril appears confident in its self-funded EagleEye project, betting that a purpose-built integrated helmet will outperform add-on attachments. As the prototypes evolve, the ultimate adoption will depend on performance in field tests, user feedback, and the Pentagon's evolving priorities.