Military acquisition cycles can last for years, which slows down tech intro.
Interaction with the wider training tech ecosystem is also a factor. For example, course authors need good VR environment editors to create usable scenarios, scoring tools, etc. They would have to be able to import their large libraries of existing vehicle and weapon models. Anecdotally, it can take an hour of scenario editing to get a minute of usable scenario
One route in would be if an existing vendor of military sims (e.g. [0]) added a VR plug-in to an existing COTS sim already used widely by the military, and the military users then got a funding opportunity (e.g. a new platform with a fresh training budget) to enhance their existing kit.
But I do understand why people don't want to work with the military.
I came from a stint at a company doing consulting and training for the defense sector mong other things. Considering the prices they asled, and got, from the government that is a big market. Adding some VR based training scenarios for the military, or emergency services, police, whatever, looks at frst glance interessting enough to at least dig deeper into. But no interesst. No idea why.