Extended reality — the umbrella over virtual, augmented and mixed reality — spent a decade as a promise. It is now quietly becoming a product category with real budgets and real users behind it. Not the mass consumer market the early hype predicted, but a set of specific, well-funded jobs where putting a screen in front of someone's eyes, or a digital layer over what they see, does something a flat display cannot.
This article is for software teams and business owners weighing an XR investment. It covers where the hardware actually stands, who is paying for XR today, what it costs to build, and how to scope a first project so that it produces a result instead of a demo.
The hardware finally caught up
Improved headset weight, higher-resolution passthrough and mainstream product launches have reduced the main adoption barriers. Meta's Quest line made standalone VR affordable and wireless. Apple Vision Pro set a new bar for display quality, eye and hand tracking, and passthrough clear enough to read a phone screen through. Android XR brought Google's platform and a second major ecosystem into the market, with Samsung and other manufacturers shipping devices. On the AR side, the phone in everyone's pocket has been a capable device for years.
For the first time, building an XR experience does not require asking users to buy niche equipment or tether themselves to a gaming PC. It does still require being honest about who owns which device. Headset ownership is a small fraction of the population, while phone ownership is nearly universal. That single fact should shape almost every scoping decision you make.
Enterprise is leading, not gaming
Training simulations, remote assistance, design review and spatial data visualisation are where XR is generating measurable return today. The flashy consumer demos get the attention, but the invoices are being paid by industry:
- Training. Procedures that are dangerous, expensive or rare to practise in reality — surgical steps, emergency response, operating heavy machinery, handling hazardous materials — can be rehearsed in VR as many times as needed. The cost of a mistake drops to zero.
- Remote assistance. An expert sees what a field technician sees and annotates the live view. One specialist can support many sites without travelling, and the fix happens on the first visit.
- Design and engineering review. Walking through a building, a vehicle interior or a factory layout at full scale before anything is built catches problems that a 3D model on a monitor hides.
- Retail and product configuration. Seeing a kitchen, a car or a piece of furniture in the actual space, in the actual finish, shortens the decision and reduces returns.
- Healthcare. Surgical planning, exposure therapy, rehabilitation exercises with tracked movement, and patient education are all in clinical use rather than in pilots.
The pattern across all of these is the same. The XR experience replaces something that was previously expensive, slow or risky, and the saving is easy to put a number on. That is why the budget exists.
Where consumer XR actually is
Consumer VR is a healthy but bounded market, mostly gaming and fitness, with a hardware base that grows but has not exploded. Consumer AR on phones is enormous but shallow — used heavily inside social apps and occasionally for shopping, rarely as a standalone destination. Spatial computing on premium headsets is real and impressive and still early, with usage concentrated among developers, enthusiasts and a few professional niches.
None of this means consumer XR is a bad bet. It means the bet should be sized to the audience that exists today, with a plan for the audience that may exist later. A phone-based AR feature inside an app people already use reaches millions; a headset-only app reaches thousands. Build for the first and design so the second is possible.
What it costs, and why
XR projects are more expensive than equivalent 2D software for reasons that are worth understanding before you sign off a budget:
- 3D content. Every object, environment and animation needs to be modelled, textured, optimised and tested. Content is often the largest single line item, larger than the code.
- Performance constraints. A headset has to render two views at high frame rates on a mobile-class chip. Dropped frames are not a polish issue — they cause nausea. Optimisation is continuous work, not a final pass.
- Interaction design is still young. There are no settled conventions for menus, text entry, selection or navigation in 3D space. Every project spends time discovering what works, and some of that time is unavoidable.
- Device testing. There is no reliable simulator for comfort, tracking quality or hand interaction. You need the devices, and you need people wearing them for real sessions.
- Platform churn. SDKs, store policies and device capabilities are changing faster than on mobile. Budget for maintenance, not just launch.
The upside is that engines and tooling — Unity, Unreal, WebXR, the native SDKs from Apple and Google — have matured to the point where a competent team can ship a production XR application without inventing infrastructure. The cost has shifted from platform work to content and design, which is where it should be.
If you are considering an XR project
- Anchor it to a concrete task, not a demo. "Reduce training time for procedure X" is a project. "Explore what VR could do for us" is a budget line waiting to be cut.
- Design for short, comfortable sessions. Ten to twenty minutes is a good ceiling for VR. Anything that demands an hour will be abandoned by most users regardless of how good it is.
- Plan for a 2D fallback from day one. The person without a headset, the one who gets motion sick, the one on a tablet in a meeting — they all need a path. A fallback also makes the content reusable across channels.
- Budget for iteration. Spatial UX is still being worked out. The first version will teach you things you could not have known in advance, and the budget should assume a second version.
- Instrument everything. Session length, task completion, points of abandonment. Without data the project will be judged on the impression it makes in a boardroom demo, which is the wrong test.
Choosing the platform and the engine
The technology choice follows from the audience, not the other way round. If the users are the public, build for phones first with ARKit and ARCore, or use WebXR and a 3D viewer so the experience opens from a link. If the users are staff with company-issued headsets, a standalone Quest or Android XR device is the practical choice, and Unity is the engine most teams reach for because of its device coverage and asset pipeline. Unreal earns its place where visual fidelity is the point — architectural walkthroughs, automotive configurators. Apple's visionOS is worth targeting when the audience is professional and the budget covers premium hardware.
Whatever the engine, keep the content pipeline separate from the application. Models, textures and animations stored in open formats such as glTF and USD can be reused across engines and devices. That separation is what keeps a project portable when the hardware landscape shifts again — and it will.
A sensible sequence
The teams that succeed with XR tend to follow the same order. They start with a phone-based AR or a web-delivered experience because it reaches the most people at the lowest cost. They measure whether the moment is real — whether anyone uses it, whether the metric they cared about moved. Only then do they invest in a headset experience, and they do it for the audience that has demonstrated the demand, usually a professional one.
They also build assets properly from the start. Optimised, well-structured 3D content built for a phone experience carries over to a headset. Content built quickly for a demo does not, and has to be redone. Reusing the same 3D assets across multiple XR experiences reduces overall project cost.
Frequently asked questions
What is the difference between VR, AR and mixed reality?
Virtual reality replaces what you see with a fully digital environment. Augmented reality adds digital content on top of the real world, usually through a phone camera or glasses. Mixed reality is AR where the digital content understands and interacts with the physical space — objects sit on real tables and hide behind real walls. Extended reality is the umbrella term covering all three.
Do we need a headset to start an XR project?
No. Most successful first projects are phone-based AR or web-delivered 3D, because that is where the users are. A headset version is a second step, justified by evidence from the first.
How long does an XR project take?
A focused phone AR feature can ship in weeks. A VR training module with custom environments and tracked interactions is typically several months, with content creation the longest phase. Anything involving custom hardware integration or multi-user sessions takes longer and should be planned in stages.
XR is no longer speculative, but it rewards teams who scope tightly and build for a real job to be done. That is exactly the kind of problem we like to help shape — from the first phone-based prototype to the production system that earns its budget.