The primary engineering challenge of VR is motion sickness caused by a mismatch of visual and inner ear information, which is extremely well established as causing people to throw up in a wide range of contexts outside of VR. The principles behind modern VR tech are decades old - the reason we're getting consumer VR now, in the late 2010s, is because commodity computing hardware is now fast enough to render scenes and track heads with a low enough latency to prevent motion sickness. However, low hardware latency is necessary but not sufficient to prevent motion sickness - the application developer also has a responsibility not to throw the user around in virtual space without corresponding real-world movement. A great many VR applications do not respect this principle - two such programs were mentioned by JabavuAdams at the top of the thread. Hence, the primary cause of nausea on VR is motion sickness caused by content.

Yes, even in a decent VR experience you will get headaches after an hour or two, possibly or even probably from vergence-accomodation conflict. But that's entirely secondary - you won't last more than a couple minutes on a virtual roller coaster.