While the consumer metaverse flopped, enterprise AR kept growing quietly. Apple Vision Pro brought spatial computing back to the headlines, but the real story started years earlier on factory floors.
When Apple announced Vision Pro in June 2023 and shipped it in February 2024, the coverage treated spatial computing like something new. The price tag (starting at $3,499) and the polished industrial design made it feel like the opening act of a new era. But for anyone watching enterprise technology closely, the reaction was closer to "finally the consumer press noticed." Enterprise AR had been accumulating real use cases for years before Apple put spatial computing on the cover of every tech publication.
Microsoft's HoloLens had been shipping to enterprise customers since 2016. The HoloLens 2, released in 2019, found traction specifically in manufacturing, surgery, and field service. Boeing has publicly described using HoloLens for wiring harness assembly, where technicians overlay digital schematics on physical aircraft components rather than consulting paper manuals. Surgeons have used it to overlay patient imaging data during procedures. Field service engineers at companies like Thyssenkrupp used it to provide remote expert assistance to technicians on-site. None of this involved the consumer metaverse. None of it required you to own a headset at home. It was a focused enterprise workflow tool solving a specific problem in a specific physical context.
Google Glass had the inverse trajectory. Launched as a consumer product in 2013 with tremendous hype, it was pulled from the consumer market in 2015 after failing to find a value proposition that outweighed the social awkwardness and privacy concerns of wearing a camera on your face in public. But Google Glass Enterprise Edition, launched quietly in 2017 and updated in 2019, found the same kind of home that HoloLens found: hands-free workflow support in manufacturing and logistics. GE Aviation used it for manufacturing operations. DHL used it for warehouse picking. The consumer product failed. The enterprise version found a niche and stayed there. The lesson was not that Glass was a bad product. It was that the use case was always industrial, not social.
Meta Quest for enterprise tells a similar story. The consumer Quest headsets were repackaged and positioned for enterprise training in 2022 and 2023. The training use case is genuinely strong for AR and VR: high-stakes procedures, equipment that is too expensive or dangerous for trainees to practice on physically, geographically distributed workforces who need standardized instruction. Walmart reportedly trained store associates with VR before the 2017 holiday season and later expanded the program. The military and various first responder agencies have used VR training for years. These are real deployments with real outcomes, but they do not generate the kind of cultural buzz that comes with a $3,499 headset running spatial video of your living room.
Apple Vision Pro is interesting to me as an IS researcher less for what it does today and more for what it signals. The device is impressive as engineering, but at launch it was primarily useful for solo productivity and media consumption, not the kind of collaborative, situated, hands-on work that made HoloLens and Glass Enterprise Edition actually valuable. The enterprise applications for Vision Pro are coming, but the hardware runs hot, weighs heavily, and has a battery life that does not fit most full-shift factory or surgical applications. What Apple did accomplish was forcing the industry to take spatial computing seriously as a category. The investment that flows into enterprise AR development because of Vision Pro will likely produce things more useful for enterprise workflows than Vision Pro itself in its first generation.
From an IS research perspective, spatial computing raises questions that the standard technology adoption literature is not well-positioned to answer. The Technology Acceptance Model, which I spend a lot of time with in my comps prep, was built around the relationship between a user and a software system. Perceived ease of use and perceived usefulness are meaningful constructs when the system sits on a screen you interact with deliberately. Spatial computing dissolves that boundary. The system is overlaid on the physical environment. Interaction is gestural, voice-based, and eye-tracked. The concept of "using" the system blurs into "being in a space where the system is active." That is a different kind of adoption question.
The data generated by spatial computing also creates IS governance challenges that most organizations have not confronted. A warehouse worker wearing AR glasses generates a continuous stream of first-person video from their working environment. That stream contains other people's faces, proprietary process information, and biometric data like gaze patterns that reveal where attention is directed and for how long. Gaze tracking is particularly sensitive. Where someone looks is a behavioral signal. It can be used to evaluate worker performance, attention, and compliance with documented procedures. It can also be used to surveil workers in ways that go well beyond what any prior workplace technology enabled. The governance question is not just technical. It is about what norms, policies, and legal frameworks govern the collection and use of that data, and who in the organization has the authority to set those norms.
The ergonomic challenge is real and underreported. Both HoloLens 2 and Vision Pro are heavy enough that extended wear causes neck fatigue. Factory workers wearing a headset for eight hours face different ergonomic demands than a developer who puts one on for a product demo. The form factor is improving across generations, but it is still a physical constraint on enterprise adoption in sustained-wear contexts. Any IS researcher studying AR adoption in manufacturing needs to account for physical fatigue as a variable, not just attitudes, self-efficacy, and intention to use.
The consumer metaverse story is worth separating clearly from the enterprise AR story because they are fundamentally different value propositions. The consumer metaverse as Meta conceived it was about creating a persistent social virtual environment where people would spend time, socialize, work, and spend money. The value proposition required enough people to be present simultaneously to create the social experience. It ran directly into the cold-start problem: the metaverse is only valuable if other people are there, and people are only there if it is valuable. Meta spent billions building the infrastructure before there was enough social mass to make the experience compelling. Enterprise AR does not have this problem. A single technician wearing a HoloLens on a factory floor gets immediate value from overlaid assembly instructions without anyone else needing to be present. The value proposition is individual and task-specific.
That distinction matters for IS research because it means the two phenomena require different theoretical lenses. Consumer metaverse adoption (to the extent it happens) is a network effects problem with strong platform dynamics and social influence mechanisms. Enterprise AR adoption is closer to a task-technology fit problem: does the technology support the specific physical work being done well enough to justify the cost and the learning curve? The failure of the consumer metaverse does not tell us anything useful about the future of enterprise AR, and treating them as the same phenomenon muddies both analyses.
What I keep coming back to is how quietly enterprise AR has been growing while the consumer press cycles through hype and disillusionment on the flashier products. The factory floor does not generate conference keynotes or breathless product reviews. But that is where the durable adoption is happening, and it is where the most interesting IS questions about situated cognition, organizational routines, and data governance are actually playing out.
About the author
Share
More notes
Related notes