What Mixed Reality is, and what it is not
Mixed Reality places digital content into the real world and gives that content an understanding of the room, so it can sit on surfaces, hide behind objects and stay where it was left.
The understanding is the whole definition. An Augmented Reality object anchored to a point will float through a table if the user walks around it. A Mixed Reality object knows the table is there. That difference sounds cosmetic and is not, because producing it requires the device to build and maintain a model of the space, and requires the studio to write content that reacts to a room it has never seen.
Mixed Reality is not a headset category. Modern devices with passthrough cameras run fully virtual scenes and room-aware ones on the same hardware, so the device no longer decides what a project is. It is also not a synonym for spatial computing, which is a platform term rather than a description of what the content does.
The practical test is simple. If the experience would still work with the real room replaced by a blank space, the project is Virtual Reality and should be priced as such. If it would still work with the digital content floating at a fixed point, it is Augmented Reality. Mixed Reality is what remains when the content genuinely has to negotiate with the physical space.
What hardware ships today
Mixed Reality requires a device that can see and model the room, which narrows the field considerably and pushes most work onto headsets rather than phones.
Passthrough headsets are the mainstream option. They present the real world through outward-facing cameras and composite digital content into that view, with enough spatial understanding to place content against real surfaces. Optical see-through devices, where the user looks through a transparent display, preserve a natural view of the room and hands but constrain field of view and brightness. The two feel materially different to wear, and a pilot that skips hands-on comparison usually regrets it.
Phone-based Mixed Reality exists and is genuinely capable of surface detection and occlusion, but holding a phone up occupies a hand and limits sessions to a few minutes. That is fine for placing furniture and unworkable for a two-handed task, which is why industrial work concentrates on headsets.
List prices are not quoted here. The hardware in this category turns over quickly, prices differ by region and manufacturers change them without notice, so a figure on a register page misleads a buyer building a business case. Price the exact configuration with the manufacturer at the time of the build.
What enterprises actually deploy
Published Mixed Reality work clusters in demonstration and brand experience rather than in the industrial guidance the category is usually sold on.
That is what the register supports. Reading across the studios currently on record, the work described as Mixed Reality is weighted toward events, showrooms and location-based experiences, where the real environment is part of the appeal and the audience is guided through a short session.
Draw & Code publishes an automotive experience for Toyota and a live performance built with Meta and Dolby.io. Magnopus documents Cryptic Cabinet, a room-aware piece built for Meta that adapts to the space it is played in, which is close to a reference implementation of what the category means. WE/AR Studio publishes a Mixed Reality rehabilitation system, one of the few clinical examples on record.
The honest reading is that Mixed Reality has fewer published deployments than either neighbouring category, and that the industrial maintenance and remote-guidance narrative is thinner in public evidence than in vendor material. A buyer scoping that work should ask for delivered examples specifically, not capability claims, because the gap between the two is unusually wide here.
What drives the cost of a build
No credible public source publishes Mixed Reality project costs, and this register does not publish one either.
Directory price data is self-reported through a submission form, which makes it a marketing figure rather than an observation. What can be stated is why Mixed Reality carries a premium over Augmented Reality for what looks like a similar result.
Three things drive that premium, and one drives the base cost.
- Unknown environments. Content must work in a room nobody on the team has seen. Every assumption about floor space, lighting and clutter becomes a failure mode, and the testing burden multiplies accordingly.
- Spatial correctness. Occlusion, physics against real surfaces and persistence between sessions each require work that has no Augmented Reality equivalent, and they are the features users notice immediately when they are wrong.
- Session logistics. Headset work implies fitting, hygiene, guidance and a supervised session. Those costs are operational rather than engineering, and they are usually missing from the first budget.
- Asset production, which behaves as it does everywhere: three-dimensional models are the largest line, and existing CAD rarely converts cleanly into something a headset renders in real time.
Ask a studio to explain which of those dominates its estimate, and ask specifically how it tests in rooms it has not seen. The answer separates studios that have shipped from studios that have demonstrated.
How long a build takes
Timeline is governed by environment testing and by device logistics, not by application development.
A Mixed Reality build is not finished when it works in the studio. It is finished when it works in a room with different dimensions, different lighting and furniture in the wrong places, and that testing cannot be compressed by adding developers. Access to representative spaces is the constraint, and it is an organizational problem rather than a technical one.
Sequence around that. Identify the real spaces early and get access booked, because a pilot in the client's actual environment is what turns a demonstration into a deployment. Device procurement belongs on the critical path too, since enterprise headset orders are not always quick and a project waiting on hardware is a project not being tested.
The register does not hold delivery timelines, because studios do not publish them consistently enough to compare.
How to assess a studio
Ask for delivered work that is genuinely room-aware, then check whether the case study describes the room doing anything.
This is the category's specific filter. A great deal of work marketed as Mixed Reality is an anchored overlay that would behave identically in an empty warehouse. A case study that describes content responding to a real surface, occluding correctly or persisting between sessions is describing Mixed Reality. One that describes a floating interface is describing something cheaper, and a buyer paying the premium should know which they are buying. The verification standard this register applies is published in full so that judgment can be made independently.
Two further checks matter here more than elsewhere.
Ask which device the delivered work shipped on, and whether it was passthrough or optical see-through. The two are different engineering targets, and experience on one transfers only partly to the other.
Ask what happens when the room is wrong. Every deployed Mixed Reality application meets a space that breaks an assumption, and a studio that has shipped will describe its fallback behaviour without being prompted.
What the published record does not tell you
Mixed Reality is the one category in this register where studios routinely name the hardware, and the reason is that the work does not run anywhere else.
Around three in five Mixed Reality case studies name a target device, against roughly one in nine across Augmented Reality work. Passthrough behaviour, hand tracking and spatial mapping differ enough between headsets that a build is rarely portable, so the device is part of the description rather than an implementation detail a studio can omit. Dates remain absent almost entirely, which matters more here than elsewhere because the hardware has turned over repeatedly.
That combination should shape how a portfolio is read. A Mixed Reality case study naming a device that is no longer sold documents real capability on a platform a buyer cannot deploy, and the studio has not misrepresented anything by publishing it. Ask when the work shipped and what the team has delivered on currently available hardware, because the answer is genuinely unknowable from the published record.
Further published work in this category
The published Mixed Reality record spans industrial visualization, live events and instruction, and it names devices across several generations.
Draw and Code publishes a stand for Philips Displays built on Magic Leap and an immersive roadshow for Mercedes-Benz, both of which place the headset in a setting the studio does not control. Event and trade show deployment carries constraints that an office pilot does not, since the hardware has to survive repeated first-time users and a fixed stand time.
Industrial and instructional work forms the other group. Treeview publishes digital twin data visualization built with Microsoft on HoloLens, and WE/AR Studio publishes Mixed Reality training for maritime students. Bit Space Development publishes The Grid, a near miss safety application targeting Pico. The common thread is that the trainee needs to see the real workspace while the overlay runs, which is the argument for passthrough over a fully enclosed headset.
Where these projects fail
Mixed Reality projects fail most often because the hardware moved underneath them.
This category has turned over faster than any other the register documents. Devices have been discontinued, platform support has ended and development stacks have been replaced, all inside the life of a normal enterprise procurement cycle. A build targeted at a headset that stops being sold is not a failure of the studio or of the software, and it is a sunk cost either way. Ask what the migration path looks like before signing, because on current evidence the question is not hypothetical.
The second failure is environmental. Passthrough and spatial mapping depend on the room. Reflective surfaces, low light, large featureless spaces and moving machinery all degrade tracking, and an industrial setting frequently contains all four. Testing in the actual space early is the only reliable check, and it is the step most often skipped because the space is in production.
Third, the use case does not need the real world. Mixed Reality costs more to build and to maintain than a fully enclosed equivalent, so it earns its place only where the trainee or worker genuinely has to see the physical environment. Where that is not true, a Virtual Reality build usually delivers the same outcome for less.
A fourth check is availability rather than capability. Before a device is specified, confirm it is still sold in every country the deployment covers, that enterprise support is offered there and that units can be replaced on a timescale the programme can absorb. Mixed Reality hardware has repeatedly been available in fewer markets than a multinational rollout needs, and a device that cannot be bought in one region turns a single programme into two.
Where to go next
Every studio named here is on the public record with its sources and its verification date. The complete list of studios delivering this work is the Mixed Reality development register.
Rankings are scoped rather than global, because a studio strong in location-based experience is not thereby strong in clinical deployment. Mixed Reality is also the category where a buyer most often arrives having been sold the wrong one, so the cheapest useful step is to test the project against the definition above before requesting proposals. Mixed Reality is the smallest of the three technology pillars by supplier count, and every studio in it also appears in at least one of the other two. Most deliver Augmented Reality development and most deliver Virtual Reality development, which is why a Mixed Reality brief is usually answered by a studio that did not start there.
Training and simulation accounts for most of the published Mixed Reality work the register documents, and it is the use case where passthrough hardware has the clearest argument over a fully enclosed headset.
