IVAS: the Army's $22 billion AR goggle ceiling, and the...

The headline number attached to the Army's augmented-reality goggle program is close to $22 billion, and it is almost always used wrong. That figure is a contract ceiling, a maximum the Army could spend over roughly ten years if everything went to plan. It is not money that left the Treasury. When the Government Accountability Office finally added up what the Integrated Visual Augmentation System (IVAS) had actually cost through 2026, the number was about $1.8 billion, and the roughly 10,000 headsets built with it were headed to storage rather than to soldiers.

Both facts matter, and keeping them apart is the whole point of this entry. IVAS is a real oversight case: soldiers got physically sick using the early hardware, a Defense Department watchdog warned the Army was buying before it knew what "acceptable" meant, and testing discipline lapsed at least once. It is also a case where the process that surfaced those problems worked, the program was restructured rather than abandoned, and the loss, while real, was a fraction of the number people quote. This post walks the record from primary sources and keeps the three separate money figures separate.

What IVAS was supposed to be

IVAS is a soldier-worn augmented-reality heads-up display, built on Microsoft's HoloLens 2 and adapted for combat use. According to the Congressional Research Service, the goal was to overlay 3D maps, sensor feeds, night and thermal vision, and targeting data directly onto an individual soldier's view, improving navigation, rapid target acquisition, and situational awareness for the Army's Close Combat Force.

The ambition is easy to understand. A dismounted infantry soldier today carries separate devices for night vision, thermal imaging, navigation, and communications, and stitches them together mentally under stress. Fusing those functions into one worn display, and doing it on adapted commercial hardware to move quickly, was a reasonable bet on paper. The question the program never cleanly answered up front was whether the resulting device would be something soldiers could actually wear and fight in for hours at a time.

The timeline

November 2018: The Army awards Microsoft a prototype Other Transaction Agreement to develop IVAS. This is a development deal, not the large production contract that came later, and the two are frequently conflated.
March 26, 2021: The Army and Microsoft sign a fixed-price production agreement with a ceiling of up to about $21.88 billion over a ten-year period of performance (a five-year base plus a five-year option), to deliver as many as roughly 120,000 devices, per Breaking Defense.
April 2022: The Defense Department Inspector General releases DODIG-2022-085, finding the Army had not defined minimum user-acceptance levels for IVAS.
May and June 2022: Operational testing (IVAS 1.0) produces the widespread physical-symptom and combat-effectiveness findings later documented by the Director, Operational Test and Evaluation and by GAO.
2022 and 2023: The Army procures roughly 5,000 IVAS 1.0 units, then roughly 5,000 IVAS 1.1 units with an improved low-light sensor.
2024 into 2025: The program reworks into IVAS 1.2, a redesigned version with a hinged flip-up flat display and a narrower field of view.
February to April 2025: Microsoft announces a plan to transfer the contract to Anduril Industries via novation; the transfer takes effect around April 10, 2025. The Army also launches a follow-on effort, Soldier Borne Mission Command (SBMC), formerly called "IVAS Next."
June 9, 2026: GAO publishes GAO-26-109135, tallying about $1.8 billion invested since 2018 and confirming roughly 10,000 headsets bound for storage.

The money, kept straight

Three dollar figures dominate this story, and blurring them is how the record gets distorted. They are not interchangeable.

The contract ceiling: up to about $21.88 billion. This is the number that produces "$22 billion" headlines. Breaking Defense reported the March 2021 production agreement as worth up to roughly $21.88 billion over ten years for as many as 120,000 devices. A ceiling is a maximum potential value, a limit on what could be obligated across the full run if the Army bought the full quantity over the full term. It is not an amount obligated, and it is not an amount spent. No serious reading of the record supports "$22 billion wasted" or "$22 billion spent."

The Inspector General's warning: "wasting up to $21.88 billion." In April 2022, DODIG-2022-085 found the Army had not defined minimum user-acceptance levels for IVAS and warned that procuring the system without attaining user acceptance "could result in wasting up to $21.88 billion in taxpayer funds to field a system that Soldiers may not want to use, or use as intended." Read carefully, this is the same ceiling number restated as risk. The IG was describing the maximum exposure if the Army pushed the full program to fielding and soldiers rejected it. It is a warning about what was at stake, not an accounting of money already gone. The report itself is partially redacted.

The money actually spent: about $1.8 billion. In June 2026, GAO-26-109135 found the Army had invested roughly $1.8 billion (close to $2 billion) in IVAS since 2018, across three acquisition efforts. This is the figure that reflects real expenditure. It is what the roughly 10,000 IVAS 1.0 and 1.1 units cost, plus the associated development, and it is what GAO says will not translate into fielded capability, because those units fell short of soldier needs and are going into storage.

There is also a fourth, smaller number worth naming so it does not get folded into the others: the FY2025 budget requested about $255 million for 3,162 IVAS 1.2 heads-up displays. That is a specific budget line for the reworked version, not part of the $1.8 billion already spent and not the ceiling.

So the honest sentence is this: the program carried a ceiling near $22 billion, the IG warned that full amount was at risk if soldiers rejected the system, and the Army actually spent about $1.8 billion before restructuring. Those are three different things.

What went wrong with the hardware

The early hardware made soldiers sick. In the FY2022 operational demonstration, GAO later reported that more than 80 percent of soldiers reported physical symptoms within three hours of use. The symptom list documented through the Director, Operational Test and Evaluation FY2022 assessment included disorientation, dizziness, eyestrain, headaches, motion sickness and nausea, neck strain, and tunnel vision. These findings pertain to IVAS 1.0, the baseline version.

The problem was not only comfort. In the June 2022 operational test, soldiers were more effective with the equipment they already carried than with IVAS 1.0: they hit fewer targets and engaged more slowly while wearing the system, and DOT&E found the system still suffered too many failures of essential functions. A device meant to make soldiers deadlier and more aware was, in that test, making them slower and less accurate.

The display also had a visibility problem in the literal sense. The active display emitted a glow that testers said could reveal a soldier's position. One tester reportedly said the emitted light "would have gotten us killed" in combat. That specific quote traces to a leaked Army report reviewed by Business Insider rather than to a named public primary document, so it should be treated as press-sourced rather than as a citation from an official release. The underlying concern, that an emitted glow can betray position, is consistent with the documented testing record.

GAO's 2026 report identified a structural root cause rather than a tuning error. The wide, roughly 70-degree field of view required a waveguide display architecture that introduced latency between a soldier's head movement and the corresponding update on the display. That lag creates a vestibular conflict, the mismatch between what the inner ear feels and what the eyes see, which is a well-understood driver of motion sickness. In other words, the cybersickness was baked into the optical design, not something a software patch could fully resolve. The IVAS 1.2 redesign narrowed the field of view from 70 to 60 degrees, part of the effort to address exactly this.

GAO also found a process lapse worth flagging plainly: after the negative soldier feedback on version 1.0, the Army did not conduct operational testing on version 1.1 before producing it. That paraphrase comes from press coverage of the GAO report, and the exact wording should be confirmed against the GAO document before being quoted as such, but the thrust is that a version went into production without the operational test that might have caught its shortfalls.

The critique and the defense, side by side

The honest critique. IVAS is close to a textbook acquisition-process failure. The Army committed to a production contract with a ceiling near $21.88 billion and pushed toward fielding before defining what "soldier acceptance" even meant, which is precisely the gap the DoD IG flagged in April 2022. The hardware that resulted made more than 80 percent of test soldiers physically ill within three hours, emitted light testers said could reveal their position, and in the June 2022 test left soldiers less effective than with their existing gear. Testing discipline then slipped further when the Army produced version 1.1 without operational testing. Seven-plus years and about $1.8 billion later, the roughly 10,000 goggles built are headed to storage rather than the field. And the core defect was designed in: the wide field of view demanded a waveguide architecture whose latency caused the sickness, so this was structural, not a fixable quirk.

The honest defense. The underlying goal is legitimate and enduring. Giving an individual soldier a fused heads-up overlay of maps, night and thermal vision, sensor feeds, and targeting data is a genuine capability advantage, and attempting it on adapted commercial AR to move fast was a defensible bet. The widely quoted "$22 billion" is a ceiling, not a loss; the actual investment was on the order of $1.8 billion, so "$22 billion wasted" overstates the damage by roughly an order of magnitude. And the Army did not keep throwing money after a failing design. It restructured into the improved IVAS 1.2, subjected the effort to further assessment, transferred program leadership from Microsoft to Anduril in 2025, and stood up a follow-on program, Soldier Borne Mission Command. The oversight system, the DoD IG, DOT&E, and GAO, surfaced the problems and forced a course correction rather than allowing the Army to field a system soldiers would refuse to use. The roughly 10,000 stored units can still support testing and development.

Both verdicts are true at once. There was real waste and real, documented harm to soldiers in testing. There was also a legitimate capability goal and a program that was restructured, not scrapped, once the evidence came in. The clean lesson is narrow and expensive: buy to a defined user-acceptance standard, tested early, before committing to production quantities. It is not evidence that the mission of putting better information in front of a close-combat soldier was misguided.

One housekeeping note for anyone researching this program: a token sometimes labeled "DASNet" does not appear anywhere in the IVAS record across CRS, the DoD IG, GAO, Army sources, or trade press. It has no established connection to IVAS and is not treated as part of this program here.

Fact-check notes and sources

IVAS is the Army's HoloLens-2-based soldier-worn AR heads-up display for navigation, target acquisition, and situational awareness for the Close Combat Force: Congressional Research Service, IF13022.
The March 26, 2021 fixed-price production agreement carried a ceiling of up to about $21.88 billion over ten years (five-year base plus five-year option) for as many as roughly 120,000 devices; this is a ceiling, not money spent: Breaking Defense.
DODIG-2022-085 (April 2022) found the Army did not define minimum user-acceptance levels and warned procurement "could result in wasting up to $21.88 billion in taxpayer funds to field a system that Soldiers may not want to use"; the figure is the ceiling restated as risk, and the report is partially redacted: DoD Office of Inspector General, DODIG-2022-085 and the full PDF.
FY2022 operational testing symptom list (nausea, headaches, eyestrain, neck strain, dizziness, tunnel vision, disorientation), with onset under three hours, for IVAS 1.0: DOT&E FY2022 assessment via Task & Purpose.
Soldiers were less effective with IVAS 1.0 than with legacy equipment in the June 2022 operational test, and the system had too many failures of essential functions: DOT&E FY2022 via Task & Purpose.
The "more than 80 percent of soldiers reported physical symptoms within three hours" figure, the roughly $1.8 billion actually invested since 2018 across three efforts, the roughly 10,000 IVAS 1.0/1.1 units going to storage, and the waveguide/field-of-view latency root cause: GAO-26-109135, June 2026, with coverage at The Defense Post and Military Times.
The display glow that testers said could reveal a soldier's position, and the "would have gotten us killed" remark, trace to a leaked Army report reviewed by Business Insider and should be treated as press-sourced: Popular Science.
Version progression (1.0 in 2022, 1.1 in 2023, redesigned 1.2 with flip-up display and 60-degree field of view), the roughly $255 million FY2025 request for 3,162 IVAS 1.2 heads-up displays, the Microsoft-to-Anduril novation effective around April 10, 2025, and the SBMC follow-on: CRS IF13022 and Breaking Defense.

IVAS: the Army's $22 billion AR goggle ceiling, and the $1.8 billion it actually spent