The U.S. Army has taken another step toward turning helmet-mounted night vision into a battlefield information system that does more than show images. Anduril Industries this week announced a $159 million prototyping award to develop a Soldier Borne Mission Command system that merges night vision, augmented reality, and edge AI into a single wearable for front-line troops. The award is one element of a larger set of prototype investments that together total more than $350 million and set the stage for competing approaches to the Army’s next-generation combat headset.
The problem set is straightforward and urgent. Night vision goggles and existing helmet-mounted systems give soldiers sight in darkness; they do not directly give perception, fused situational awareness, or tools to command robotic teammates. Tactical decision making in contested, communications-degraded environments often forces squad leaders to juggle maps, radios, and separate apps just to build a common picture. Delays in assembling that picture cost time, and in combat time means lives. The Army’s doctrine is blunt about this reality: modern mission command requires forces to make and implement decisions faster than the enemy. The SBMC effort aims to collapse the time between observation and action by converging sensors, data, and command functions into the soldier’s head-worn display.
Anduril’s approach centers on software, modular hardware, and human-centered design. The company will build SBMC-A, the software architecture that integrates helmet displays, body- and vehicle-borne edge compute, and a partner ecosystem of sensor and software vendors. Anduril’s Lattice platform will act as the mesh backbone for SOF-esque workflows, connecting third-party data sources and enabling soldiers to task unmanned systems directly from their HUD. The firm says it has already integrated IVAS 1.2 headsets as surrogates and logged more than 260,000 hours of soldier input from prior IVAS iterations to tune the user experience and requirements. Those user inputs have been validated in multiple soldier exercises and combat training scenarios.
Technical ambition is high. The SBMC prototype combines advanced night vision with mixed reality overlays that fuse day, night, and thermal imagery with geospatial intelligence, friendly force locations, and AI-assisted cues. The system is being developed with a collection of commercial partners selected for specific domains: Meta on optical and display integration, Qualcomm on compute and silicon, Palantir and Maxar on data fusion and geospatial feeds, and Gentex on helmets and ergonomics, among others. That partner blend mirrors a wider trend where commercial optical, chip and software capabilities are folded into defense systems to accelerate performance and lower development risk.
A practical capability demonstrated in recent field tests was the ability to task and control drones from a soldier-worn headset. Anduril reports successful demonstrations where operators used Lattice-integrated headsets to task unmanned aerial systems at distances beyond three kilometres via line-of-sight radio links, without a dedicated drone pilot. That example is important because it shows how SBMC seeks to reassign roles and responsibilities: the sensor-to-shooter loop grows shorter when a single operator can both perceive and influence the battlespace from a forward position.
Speed of software delivery is another design focus. Anduril claims it cut over-the-air software update timelines from two days to about 15 minutes for fleet devices by using optimized test and fleet management tools in Lattice. Rapid patching and iterative updates are crucial when fielded systems must adapt to changing threats, radio environments, and user feedback; they also create new demands for secure update processes and robust key management. Faster updates lower the friction between operational feedback and system improvement, but they also demand rigorous cybersecurity controls and transparent supply chain assurances.
The award sits inside a complicated program history. The Integrated Visual Augmentation System, or IVAS, was an earlier Army effort that invested heavily in head-worn displays and collected extensive soldier feedback while encountering production, ergonomics, and policy challenges. The SBMC program is both an evolution and a reframe: it aims for modular, swappable hardware and an open software architecture rather than a monolithic, single-vendor headset. The Army’s reallocation and prototyping choices reflect that shift toward competition and rapid iteration; one observable result is the parallel award to Rivet Industries, a Palantir-backed team, which received a larger prototype contract and will compete in the same developmental space. Dual-path prototyping is an explicit attempt to mitigate program risk and keep multiple technical approaches in play.
Operationalising these systems presents nontrivial constraints. Disconnected or degraded communications, electromagnetic interference, and hostile cyber activity remain everyday realities for deployed units. Private edge compute and on-device AI reduce reliance on wide-area links; yet they also require comprehensive validation for performance under GPS denial, signal jamming, and spectrum congestion. Ergonomics and weight are practical limits as well; adding batteries, computers, sensors, and radios to helmets cannot compromise human endurance or safety. The Army and industry must therefore trade off features for wearability and resilience, and test systems under contested, signal-denied conditions to see how they perform when it matters most.
There are clear, real-world precedents for the SBMC model. The German and British armed forces have run iterative experiments with helmet-mounted sensors, drones, and private tactical networks; NATO-collaborative exercises have increasingly tested how sensor-fed situational awareness can be federated across multinational command chains. Commercial examples of rapid software delivery and edge-AI integration come from cloud-native firms and telecom providers that adapted similar patterns for enterprise and public-safety users. Those precedents do not eliminate the hard work ahead; they do show that mixed ecosystems of commercial and defense suppliers can field capabilities faster when the software architecture is open and the user feedback loop is short.
For soldiers on the ground, the promise is tangible: fewer devices to manage, faster understanding of the immediate fight, and the ability to task and receive data from unmanned systems without leaving the squad’s tactile flow. For program managers, the challenge is to translate prototype demonstrations into fieldable kits that survive supply chains, training cycles, and the rigor of combat. The Army’s strategy of funding multiple prototype pathways recognizes this friction; real advantage will be determined by how well any candidate system performs in long-duration exercises and in environments where adversaries attempt to disrupt sensing and communications.
The SBMC award to Anduril is not an endpoint; it is the beginning of an accelerated cycle of prototyping, soldier evaluation, and hardening. If the technology delivers on its stated goals—integrating sensors and AI into ergonomically credible hardware that updates securely in the field—then squad-level command will shift markedly from information assembly to information exploitation. That shift will change training, doctrine and logistics as much as it changes helmets. The next months of exercises and developmental testing will show whether this iteration of mixed reality moves from bold promise to routine capability.
