Anyone who has demoed VR publicly has seen the same face. A visitor puts on the headset, takes one look at something the designer thought was harmless, and takes the headset right back off. Motion sickness in VR is a design failure, not a user failure - and it's almost always caused by specific mistakes in the first 30 seconds of the experience.
Here's the practical checklist for VR work that doesn't make people sick, and a few principles for making the first half-minute work as hard as it needs to.
Never move the camera the user didn't ask to move
The single biggest contributor to VR motion sickness is a mismatch between what the inner ear feels (nothing, because the user is standing still) and what the eyes see (movement). The rule that follows: the camera only moves when the user moves their head. Period. No forced camera pans, no zooming, no dolly shots, no cinematic camera language that works in film.
Teleport locomotion is the standard solution when the experience needs to cover distance - blink to the next position rather than gliding. Smooth locomotion (flying, walking, vehicle movement) is acceptable for experienced users with comfort-toggle controls, but should never be the default for first-time visitors.
The first 30 seconds sets the tolerance
A visitor's motion-sickness tolerance in the first 30 seconds is their tolerance for the rest of the experience. A comfortable opening lets them acclimate; an uncomfortable opening puts them on a slow slide into nausea no matter how gentle the middle of the experience is.
Practical translation: the first 30 seconds should be stationary, visually simple, at a consistent scale, with a clear visual horizon. Give the inner ear time to commit to the virtual space. Then introduce complexity, motion, and scale changes - one at a time, never all at once.
Scale is non-intuitive
Objects in VR need to be at real-world scale if the experience wants to feel natural. Scale errors - a virtual chair that's slightly too big, a doorway that's shorter than it should be - register as uncanny in a way that's very hard to articulate but easy to feel. A visitor may say the experience felt "off" without being able to name the scale issue.
Extreme scale changes - shrinking the user to ant size, growing them to giant size - are powerful but disorienting. They need to be transitioned deliberately, with a clear anchoring moment between scales, not cut to suddenly.
Frame rate is not negotiable
VR runs at 90 frames per second for comfort. Dropping below 90 - even briefly, even for a single frame - causes immediate motion sickness that builds the longer the dips last. This means VR experiences have to run the rendering budget much tighter than a 60-fps desktop game would. Every effect, every particle system, every draw call is scrutinized against the 11-millisecond frame budget.
Motion reprojection (Oculus's ASW, SteamVR's Motion Smoothing) can bridge occasional dips, but it adds artifacts of its own and shouldn't be the default strategy. Hitting target frame rate natively is the real goal.
A beautiful 60-fps VR scene is worse than a modest 90-fps one, every time. Ship at frame rate, then scale up the art.
Accessibility beyond motion sickness
- Seated mode as a first-class option. Not every visitor can stand for a 10-minute experience.
- Height calibration respect. Tall users and short users should both have ergonomically correct perspectives, not one fixed height.
- Controller-free affordances. Not every visitor can operate a controller; voice or gaze-based input covers more of the audience.
- Breathing room in the narrative. Moments where the user isn't doing anything demanding give everyone a chance to regulate.
VR continues to be the single most visceral presence medium available. It also continues to be the one most demanding of craft in how it's deployed. The designers doing interesting work in 2026 are the ones treating comfort as a design constraint, not a debug option.
