What spatial intelligence is — and why it’s the next layer of computing

For thirty years we’ve squeezed the world into rectangles. Documents, maps, shops, conversations, entire cities — all flattened onto a screen and navigated with a cursor or a thumb. It worked remarkably well. But it quietly trained us to accept a strange trade: to get information, we look away from the space we’re standing in.

Spatial intelligence is the move back. It’s the ability of software to understand physical space — its geometry, the objects in it, the people moving through it — and to place digital information into that space, where it actually belongs. Less “pull out your phone and search,” more “the answer is already here, anchored to the thing you’re looking at.”

How it differs from AR and VR

People use “AR,” “VR,” and “spatial computing” almost interchangeably. It helps to separate the display from the intelligence:

AR / VR / XR describe how content is shown — overlaid on the real world, or replacing it entirely.
Spatial computing is the broader category of computers that operate in three dimensions.
Spatial intelligence is the layer underneath: the understanding of what a space is and how people behave in it. It’s what makes a digital tree stay anchored to a real pavement, or lets a system notice that everyone avoids a particular corner of a room.

A headset without spatial intelligence is just a screen strapped to your face. The intelligence is the part that matters.

Why now

Three curves are crossing at once. Capture is cheap — any modern phone can scan a room. Models can finally interpret what’s captured — recognising objects, surfaces, and movement in real time. And generative AI can produce convincing spatial content on demand instead of having it hand-built by a studio. Put together, they make it economically realistic to treat physical space as a programmable surface.

The interface of the next decade isn’t a screen you look at. It’s the space you’re already in.

What it changes in practice

This isn’t only a consumer-gadget story. The most interesting work is happening where space, data, and decisions meet:

Cities & architecture. Residents can preview and shape changes to a street before construction, and planners can read those choices as data. I’ve explored exactly this in AR crowdsourcing and in VR space audits.
Brands & retail. Products can be placed, tried, and understood in a customer’s own environment — closing the gap between seeing and believing.
Content & media. Generative tools like Magic collapse the cost of producing spatial and cinematic content, so good ideas aren’t gated by huge production budgets.

The honest caveats

Spatial intelligence is powerful, which is exactly why the constraints matter. Sensing space means sensing people, so privacy can’t be an afterthought — anonymised, aggregate signals beat individual tracking. And spatial design has a higher bar than flat design: get comfort, scale, or pacing wrong and people feel it in their body, not just their patience. The teams that win treat the human as the brief, not the technology.

Where to start

If you’re a founder or a team wondering where this fits, don’t begin with a headset. Begin with a decision you’re currently making with bad information — a layout, a route, a product placement, a customer journey — and ask what becomes possible once space itself is something software can read and write. That’s usually where the real value is hiding.