The Army paid $2B for headsets soldiers couldn't wear. Anduril says the fix is a software company wearing a defense contractor's skin.

What the Army Actually Bought

The U.S. Army awarded Anduril a $159 million contract in September 2025 to prototype Soldier Borne Mission Command, a helmet-mounted mixed-reality platform that fuses night vision, augmented reality, and AI-driven tactical command into one device for dismounted troops at company level and below. Rivet, a Palantir-backed startup, landed a parallel $195 million deal the same week, putting two teams against each other in a competitive prototyping phase.

SBMC replaces the Integrated Visual Augmentation System, a program that began in 2018 when Microsoft won a 10-year production contract valued at up to $22 billion to build more than 120,000 custom headsets based on its HoloLens 2. The Army spent roughly $2 billion to buy about 10,000 units across early versions 1.0 and 1.1, but most went into storage after soldiers reported headaches, nausea, eyestrain, and neck pain during field tests. Congress cut IVAS funding in 2022. The Army paused fielding, split the program into iterative versions, and tested a redesigned 1.2 model with a hinged, flat display and a 60-degree field of view (down from the original 70-degree design) that soldiers could flip up.

By early 2025, Microsoft moved to transfer the entire IVAS contract to Anduril, a company that had already integrated its Lattice AI platform into IVAS 1.2 hardware for field testing. The handover set the stage for SBMC, which the Army rebranded from "IVAS Next" and broadened beyond a single headset. SBMC now includes body-worn compute, companion wearables like watches, and an open software backbone called SBMC-Architecture built on Lattice. Anduril's partners on the program include Meta, Qualcomm, OSI, Gentex, Palantir Technologies, L3Harris Technologies, Persistent Systems, Sierra Nevada Corporation, Maxar Intelligence, and Kägwerks.

The contract covers an initial prototyping phase, not mass production. Anduril says the first scaled delivery of SBMC is slated for 2027. The Army has not committed to a single winner or a single form factor. Founder Palmer Luckey told reporters the program will likely field "dozens" of different headsets over the next decade rather than one device for every soldier, and that most SBMC hardware sold to the Army will probably come from adaptations of commercial AR and VR devices rather than from Anduril itself. The company's role is the software layer and the integration architecture, the AI backbone that fuses sensor data, blue force tracking, drone control, and navigation into one display.

The SBMC solicitation asked vendors how their systems would mitigate visual discomfort and nausea, the exact problems that nearly killed IVAS. Luckey, who founded Oculus VR before starting Anduril, said the company's existing EagleEye hardware does not cause motion sickness and that Anduril has already proven it can build head-mounted displays soldiers can use without acclimation. The Army's solicitation described SBMC as a "fight first" system with training as a secondary benefit, signaling a shift from IVAS's dual combat-and-training focus to a platform built around actual combat operations in contested, communication-degraded environments.

Why Anduril Won — and What It Means for Defense-Tech Hiring

The Army's original IVAS program was supposed to be Microsoft's. That effort, launched in 2018 and tied to the HoloLens platform, ran into a wall: soldiers reported headaches and nausea during use, and Congress cut funding in 2022. The program was nearly dead. Anduril didn't just inherit the wreckage; it convinced the Army that the failure was architectural, not conceptual. The problem wasn't mixed reality on the battlefield. The problem was that Microsoft's stack wasn't built to process real-time sensor data at the edge, on a moving soldier, in a contested environment. Anduril's pitch was that its Lattice platform already did that for drones. Putting it on a helmet was the next logical step, not a leap.

That argument carried weight because Anduril had proof. The company had already integrated Lattice into IVAS 1.2 headsets for field testing, and soldiers used the system to task drones more than three kilometers away directly from the helmet, with no dedicated operator. Software updates that once took two days can now be pushed in 15 minutes. The Army absorbed more than 260,000 hours of soldier feedback from the earlier program and folded it into SBMC's requirements.

The contract structure itself signals a shift. Anduril leads a team that includes Meta, Qualcomm, Gentex, Palantir, L3Harris, Persistent Systems, Sierra Nevada Corporation, and Maxar Intelligence. The hardware is modular (soldiers select components per mission) and the software architecture is open. That's a deliberate break from the monolithic, contractor-locked model that defined the original IVAS.

The hiring signal is already visible. Zero G Talent's board shows Anduril added 245 roles in the past 7 days, spanning Costa Mesa, El Segundo, Lexington, and Hudson, New Hampshire. The listings cut across the categories this contract demands.

• Computer vision and sensor fusion engineers, the core of Lattice's ability to merge day, night, and thermal imagery into a single coherent display
• Embedded AI and edge-computing specialists, processing sensor feeds on the helmet without routing back to a command post
• AR/VR software engineers; a Senior Software Engineer, Augmented Reality Intelligence Systems role is already posted, and it sits at the intersection of real-time rendering and battlefield data
• Human-factors and optical engineers: the original IVAS failed partly because the hardware made soldiers sick; fixing that requires people who understand how a helmet-mounted display interacts with a moving human body
• Modeling and simulation engineers: a Senior Modeling and Simulation Engineer, Space role at El Segundo suggests the simulation work extends beyond soldier systems into Anduril's broader portfolio
• Sustainment and supply chain leads: the Sustainment Lead, Advanced Effects and Director, Supply Chain roles indicate the company is building production and logistics capacity, not just prototyping

These aren't traditional defense-electronics jobs. A standard military hardware contractor hires RF engineers and systems integrators who work to a fixed spec over a multi-year development cycle. Anduril's hiring profile looks more like a consumer AR company crossed with an autonomous-vehicle startup, people who ship software weekly, train models on field data, and treat hardware as a platform rather than a product. The salary ranges reflect that competition:

Anduril's Costa Mesa headquarters houses both software and hardware engineering under one roof, alongside business development and mission operations. That co-location matters for a program like SBMC, where the gap between a software update and a fielded capability is measured in hours, not months. The soldier borne mission command contract is the first time that pitch extends from drones and undersea vehicles to the soldier's own sensory system, and the talent market is responding.

The Workforce Signal: AI + AR + Edge Computing on the Battlefield

The IVAS relaunch doesn't just rebuild a headset; it builds a hiring pipeline for a job category that barely existed in defense three years ago. Anduril's open-roles page lists 245 positions added in the past 7 days spanning Costa Mesa, and a significant share of them sit at the exact intersection this program demands: computer vision, sensor fusion, embedded AI, and perception engineering running on hardware that has to work when the soldier wearing it is moving, shooting, and cannot afford a reboot.

Look at what the company is actually hiring for. A Senior Robotics Software Engineer for Sensor Fusion in Washington, DC (salary range $191,000 to $253,000) writes the algorithms that ingest raw data from multiple onboard sensors and fuse them into a single real-time picture. The posting asks for Kalman filtering, state estimation, C++ or Rust, and the ability to run code on Linux-based compute modules mounted on physical systems. A Computer Vision Engineer in Costa Mesa ($166,000 to $220,000) handles camera integration across RGB, thermal, fisheye, and event-based sensors, plus the calibration and geometric vision pipelines that make those feeds usable. These are not the roles Lockheed Martin was posting for its electronic warfare programs five years ago.

The difference is architectural. Traditional defense-electronics jobs split cleanly: one team built the radio, another wrote the signal-processing firmware, another handled the display. Anduril's IVAS work collapses those layers. The job postings describe engineers who own the full stack from sensor to rendered overlay, someone who can write a visual odometry algorithm in the morning and debug its latency on an embedded GPU in the afternoon. The Sensor Fusion posting is explicit: "this role is 80% active development." The Computer Vision role wants someone who can calibrate a multi-camera rig in the lab and then deploy the calibration method to field units. The expectation is that one person moves between theory, code, and hardware without handing off to another team.

This is what "AI-native" means in practice for the workforce. It is not a buzzword on a slide. It is a hiring profile that did not exist in the defense primes' standard classification system, someone who writes production C++ for a wearable computer, understands camera geometry well enough to calibrate thermal-to-RGB alignment on a headset that gets dropped in mud, and can ship that code to a physical system within months. Army University Press has flagged this gap directly, noting that advanced equipment like IVAS "demands a mix of technical skills and situational awareness" that current military training pipelines were not designed to produce.

The salary numbers tell their own story. Anduril's listed ranges sit well above what traditional defense-electronics positions at the primes pay for equivalent experience levels. That premium is the market pricing a scarce skill set: people who can do real-time computer vision on embedded hardware and are willing to work on systems that go to war. Zero G Talent's board shows the company hiring across Costa Mesa, Washington DC, Lexington, and El Segundo, with roles like Camera Test Engineer ($98K–$130K) and Senior Modeling and Simulation Engineer for Space ($191K–$253K) feeding into the same sensor-and-perception pipeline that IVAS depends on.

The IVAS contract will not fill all these seats by itself. But it creates the programmatic demand that justifies the hiring, and the hiring, in turn, builds the workforce that can compete for the next soldier-side program. That is the signal.

The $61B Valuation Context: Why Investors Are Betting on Anduril's Combat-Systems Pipeline

Anduril's $159M IVAS contract doesn't exist in isolation. It sits inside a company that closed a Series H round at a $61 billion valuation, making it one of the most valuable private defense-tech firms on the planet. That number reflects a thesis that's become legible to venture investors over the past three years: soldier-side AI hardware isn't a science project anymore. It's a product category generating government revenue at scale.

The IVAS relaunch is the thesis made concrete. Before this win, Anduril's revenue story rested primarily on autonomous systems, the ALTIUS 700 loitering munition, the Ghost drone family, and a set of undersea surveillance products. Those are real programs with real deliveries, but they live in categories the Pentagon has funded for years. A mixed-reality combat headset worn by frontline soldiers occupies a different tier. It's a flagship modernization program, the kind that locks in production contracts for a decade and creates a sustainment ecosystem around it. Anduril's careers page already reflects that shift: the company added 245 roles in the past 7 days spanning sustainment leads, logistics program management, and micro-electronics technicians. You don't hire that breadth for a demo. You hire it for a production pipeline.

The $61B valuation also signals something about how investors are re-rerouting capital away from general-purpose AI plays. Large language models still consume the majority of AI venture dollars, but the returns picture is murky (enterprise software margins, uncertain enterprise adoption curves, and a compute cost structure that punishes startups). Defense-side AI hardware offers a different math: a single customer with a $900 billion addressable budget, deployment timelines that favor first movers, and products that don't depend on consumer willingness to subscribe. Anduril's pitch to its Series H backers reportedly centered on exactly this, that the company had moved past the prototype stage and into programs where the transition from contract to production contract was a matter of execution, not persuasion.

The IVAS relaunch strengthens that argument in a way that an autonomous drone contract alone could not. Head-worn systems create a data feedback loop that Anduril's other platforms don't: a soldier wearing an AI-enabled headset generates training signal for computer vision models, sensor fusion algorithms, and human-factors optimization every time they use it. That loop tightens the product, which tightens the contract, which tightens the revenue. Investors who bet on Anduril at $61B aren't buying a drone company. They're buying the argument that the next generation of defense AI revenue will flow through systems the soldier carries, not just systems the soldier launches.

That's why the IVAS win matters beyond its $159M headline figure. It's the proof point that soldier-side AI hardware can capture a program of record, the Pentagon's most durable funding category. And it's the reason Anduril's board is filling with sustainment and logistics roles alongside the modeling and simulation engineers who built the prototypes. The next phase isn't invention. It's production at scale, and the workforce to match it.

AI-Native Warfare's First Mass-Hiring Frontier

The Soldier Borne Mission Command contract is, on paper, a $159 million deal to build mixed-reality headsets for dismounted soldiers. In practice, it's the first time a company built around AI-native architecture, not bolted-on software, not a Microsoft HoloLens derivative, has captured a flagship soldier-modernization program from the U.S. Army. That distinction matters for anyone tracking where defense-tech employment is actually heading.

The original IVAS program, awarded to Microsoft in 2021, was a conventional defense-electronics effort: a hardware-heavy headset with software layered on top. Soldiers reported nausea, headaches, and eye strain during testing. The program nearly died. When the Army restructured the requirement and opened SBMC to competition in 2025, Defense One reported the service awarded more than $350 million across two teams, Anduril at $159 million and Rivet at $195 million, with a mandate that was fundamentally different from the original: AI-driven situational awareness, day/night sensor fusion, and mission command at the edge, not a glorified heads-up display.

Anduril's win is a hiring signal. Zero G Talent's data shows the company added 245 roles to its board in the past 7 days spanning Costa Mesa, Lexington, Hudson, and El Segundo. The open positions (camera test engineers, modeling and simulation leads, micro-electronics technicians, supply chain directors) read less like a traditional defense contractor's requisitions and more like a hardware-software integration shop that happens to build weapons systems. Sustainment leads sit alongside advanced-effects engineers. That mix is the workforce blueprint for AI-native warfare: people who can ship hardware, update its models in the field, and maintain it without sending it back to a depot.

Palmer Luckey has been explicit about the recruiting philosophy behind this buildout. After watching Oculus's post-acquisition hiring funnel narrow almost exclusively to Bay Area engineers, a pool he described as "mercenary-minded," he's pushed Anduril to recruit nationwide, with a specific focus on veterans. The pitch, as he's framed it, is "work that matters" over the standard Silicon Valley compensation package. Whether that pitch holds is an open question, but the application pipeline has clearly been large enough to sustain a 4,000-job expansion in Long Beach alone, a figure the city's mayor cited in 2028 employment projections tied directly to Anduril's growth.

The broader context is what makes SBMC a leading indicator rather than a one-off. Anduril closed the largest funding round in defense-tech history in June 2025, a Series H that valued the company at $61 billion. That valuation rests on a bet that soldier-side AI hardware, headsets, autonomous drones, undersea systems, the connective tissue between them, is a venture-scale category, not a niche. SBMC is the first proof point that the Army agrees. When the service's flagship wearable program goes to a company whose core product is an AI command-and-control platform, not a pair of goggles, the talent market takes notice.

For anyone working at the intersection of AI, AR, and embedded systems, the implication is concrete: the next hiring wave in defense tech won't look like Lockheed Martin's satellite division or Raytheon's radar shops. It will look like Anduril's Costa Mesa floor, computer vision engineers next to supply chain managers, firmware teams next to human-factors designers, all building systems that update their own models after deployment. The jobs are open now. The contract just confirmed the thesis.

Working in frontier tech? Zero G Talent tracks the openings: browse frontier tech jobs, openings at Anduril Industries, and the people building the field.