The architecture industry has been flirting with VR and mixed reality for a decade. Walk-through visualisations. Client presentations in virtual spaces. Design reviews with goggles on. Most of it was impressive technology in search of practical value — demos that wowed clients but didn’t fundamentally change how buildings were designed or reviewed.

Spatial computing is different. Not because the technology is radically new, but because it’s finally good enough to integrate into actual design workflows rather than existing as a parallel demonstration track.

What’s Changed

Three things converged to make spatial computing practically useful for architecture in 2026.

Passthrough mixed reality. Previous VR approaches required architects to leave their physical workspace entirely. Put on a headset, enter a virtual world, interact with a virtual model, take off the headset, return to your desk. The context switching killed productivity. Modern mixed reality headsets — the Apple Vision Pro, Meta Quest 3, and their successors — overlay digital content onto the physical workspace. You can view a 3D model at desk scale, then walk through it at full scale, without losing sight of your surroundings or your colleagues.

BIM integration. Building Information Modelling software — Revit, ArchiCAD, Vectorworks — is where architects actually work. Spatial computing tools that require exporting and converting models add friction that prevents adoption. Direct BIM integration, where changes in the spatial computing environment reflect in the BIM model and vice versa, is now available from several vendors. Arkio has been particularly effective at bridging this gap.

Collaborative spatial environments. Architecture is collaborative. Design decisions involve architects, engineers, clients, and contractors working together. Spatial computing tools now support multi-user sessions where participants — some wearing headsets, some on screens — can view and discuss the same 3D model simultaneously. This isn’t fundamentally new, but the quality and reliability have reached the point where firms trust it for client meetings.

Real Firms, Real Projects

Design review and clash detection. An Australian practice I spoke with — mid-size, about 60 staff — has been using spatial computing for design review sessions over the past eight months. Their process: the design team builds in Revit as normal, then holds weekly design reviews where the team walks through the model in mixed reality.

They’ve found spatial review catches coordination issues that 2D drawings miss. Ceiling heights that feel right on a plan but oppressive in 3D. Sightlines from specific positions that weren’t considered. Services routing conflicts between structural, mechanical, and hydraulic systems that are obvious when you can see all the systems at once but easy to miss in layered 2D drawings.

The practice estimates they’ve caught 15-20 coordination issues per project in spatial reviews that would previously have been discovered during construction — when fixing them costs 10-50 times more. Over their last four projects, they estimate the spatial review process saved $200,000 in rework costs.

Client communication. One of the most consistently cited benefits is client communication. Architects understand plans and sections. Most clients don’t. Showing a client a 3D model they can walk through, at the scale of the actual building, produces more informed feedback than any drawing set.

A residential architect told me her client approval process has shortened by 30-40% since adopting spatial presentations. “Clients used to approve plans, then panic during construction because reality didn’t match their mental image of the drawings. Now they’ve experienced the space before we build it. The surprises are gone.”

On-site construction verification. This is the emerging use case with perhaps the most long-term potential. Overlaying the BIM model onto the actual construction site — seeing where walls, penetrations, and services should be relative to what’s actually built — enables real-time verification during construction. Trimble’s XR10 has been the primary tool for this application, though adoption is still early.

The practical challenges are significant: construction sites are dusty, bright, and require safety equipment that doesn’t always coexist with headsets. Battery life limits sessions. And the models need to be positioned precisely relative to the physical structure, which requires surveying and calibration that adds setup time.

What Doesn’t Work Yet

Full design creation in spatial computing. Despite vendor demos showing architects sculpting buildings in mid-air, the reality is that spatial computing isn’t an efficient design tool for detailed work. Precision input — exact dimensions, specific material specifications, complex geometric relationships — is still faster with keyboard, mouse, and traditional CAD/BIM tools.

Spatial computing excels at review, communication, and verification. It’s a complement to traditional design tools, not a replacement. Firms that tried using it as a primary design tool found their architects slower and more frustrated than with conventional workflows.

Remote collaboration across time zones. The technology supports it, but the experience degrades significantly when participants are on different quality connections or using different hardware. A review session where some participants are in Apple Vision Pro, some on Quest 3, and some on a laptop screen produces an inconsistent experience that undermines the benefit.

Spatial collaboration works best when participants are either co-located (using headsets together) or all using similar-quality hardware on good connections. Mixed scenarios — which are the reality for most project teams — are workable but not yet smooth enough to be the default collaboration mode.

The Cost Question

An Apple Vision Pro costs $3,499 USD. A Quest 3 costs $499 USD. Software licenses for spatial design tools run $100-$500 per month per user. For a 60-person practice, equipping the design team with spatial computing tools represents a $50,000-$150,000 investment depending on hardware choice and software needs.

Against the $200,000+ in rework savings estimated by the practice I mentioned, the ROI is there — but only if the tools are actually used consistently. Several firms I spoke with bought headsets that are now gathering dust in a closet because the initial enthusiasm faded and nobody mandated their integration into standard workflows.

The firms succeeding with spatial computing have designated champions who schedule reviews, maintain models, and troubleshoot issues. Without someone driving adoption, the technology reverts to occasional demos rather than integrated practice.

Spatial computing in architecture has crossed from “interesting demo” to “useful tool.” It hasn’t yet reached “essential practice,” and it may not for another few years. But the firms investing now are building skills and workflows that will compound as the technology improves. The gap between firms that can offer spatial design review and those that can’t will become a competitive differentiator — it’s just a question of when.