Seems to work pretty well for u-he. They have one the most generous demo/open-beta policies of any any plugin dev and as far as we can all tell, business isn't taking a hit as a result. Imo time-limited demos, particularly anything less than 30 days, can encourage impulse buying, as one may feel pressure to buy, lest they lose the ability to continue working with the plugin in a project/s. Impulse buys aren't really in the best interest of devs and customers alike, at least not in the long term. A dev gets their cash whether a purchase was considered or not, but speaking from experience (I'm assuming many others have fallen foul also) plugins bought on impulse tend to get left by the wayside and eventually sold on, meaning that we the customer don't become invested in the product and are therefore less likely to purchase future upgrades, soundsets etc. As for the customer side, demoing lots of products and feeling pressured to buy them can result in a library full of plugins that you don't end up using. Once the initial excitement of a shiny new toy wears off and the pressure of project compatibility is alleviated, you're left with a bunch of tools that you didn't actually need. It's not great for your creative output nor your wallet.
It's not to say that having open-ended demos won't lead to impulse purchases, but imo if you have longer to try a plugin, to get over the honeymoon period and fully incorporate a new tool into your workflow, then you are much more likely to continue using it and investing in it's development, supporting the dev in the loner term as a result.
TLDR; I don't think you are correct to assert that infinite demos are universally/unquestionable negative for everyone involved.