Nobody in this industry publishes numbers. You ask three agencies what a VR training module costs and you get three meetings before you get a figure. I understand why, because every project genuinely is different, but it does not help somebody who has been asked to put a line in next year's budget.
So here are mine.
The short answer
A single-scenario XR training simulator is £10,500 to £15,000 and takes nine to sixteen weeks. That is one environment, one route through it, delivered to Meta Quest, Apple Vision Pro or a business headset like PICO, with the 3D content supplied by you.
The full range across everything I build is on the pricing page. Day rate is £600, half day £300, and £75 an hour for ad hoc work at the moment.
Those are freelance numbers. An agency quoting the same brief will be higher, sometimes two or three times higher, and that is not a markup for the sake of it. You are paying for account management, medical writing, a studio and somebody to answer the phone when I am on a plane. Sometimes that is worth it. On a single module it usually is not.
What actually moves the number
Four things, in roughly this order.
Platforms. One headset is one build. Two headsets is not two builds, but it is a good deal more than one and a half. Every extra target adds configuration, device testing and usually a couple of code paths that exist only because of it. If you can commit to Quest alone, do that, and add the others later if the thing lands.
Assets. This is the big one and it catches people out. A simulator is mostly 3D models, animation, voiceover and UI. If you already own production-ready 3D, and most pharma clients own more than they think, the cost drops a long way. If everything is built from a standing start, expect the asset work to be half the project. Ask whoever produced your last mode of action animation what they still have on file.
Integrations. SCORM or LMS reporting, single sign-on, a live data feed, custom analytics. Each one is a fortnight you were not planning for, and each wants scoping on its own rather than being folded into "and it reports back".
The deadline. A congress date is not really a deadline, it is a fixed wall. You cannot move it, so contingency has to go in at week one rather than get found in week ten. A hard date does not always cost more, but it does mean scope has to lock earlier than anyone finds comfortable.
What makes it cheaper
Supplied assets and a single platform, as above. Past that:
Reuse. The second scenario in the same environment is much cheaper than the first, often less than half. If you know you want three modules, scope three at the start rather than commissioning one and coming back in a year.
Starting medical and legal review early. This costs nothing and saves weeks. Review is happening either way, and the difference between finding a problem in week three and finding it in week fourteen is the whole schedule.
Deciding what you are not doing. Most briefs I see have one feature in them that accounts for a quarter of the build and almost none of the value. It is usually multiplayer.
What a fixed price covers
Mine covers the build, two structured review rounds, QA on the target hardware and a post-launch support window. It does not cover headsets, on-site support at the event, or content written from scratch. Those get quoted separately, because bundling them hides what they actually cost you.
When not to build it
VR is the wrong answer more often than the people selling it tend to admit.
If the content is linear and the user makes no decisions, you want a film. It will look better and cost less. If it has to reach a thousand reps next month, you want web or an iPad app, because the logistics of headsets at that scale are a project in their own right. And if the budget is under about £8,000, you will not get something that holds up next to whatever the delegate saw on the stand opposite.
Where it earns the money is narrower. When the user has to do something in a particular order, when getting it wrong in real life is expensive, or when the thing only makes sense at a scale you cannot photograph. Procedural training, mechanism of action, and anything where somebody needs to feel the size of the space they are standing in.
Getting a number for your own project
Send me the brief. I will come back inside 48 hours with a scope and a fixed price, and if I think the money is better spent on something that is not VR, I will say so.
There is more on how these get built on the XR simulators page, and the case studies show what has come out of them.