I refused to do that when I did induction and LED lighting. My products are still in use by the people that bought them. I don't expect them to need a replacement for 20 years. Maybe by then I'll be back in the lighting business.
Were I making such superior LEDs, I'd market the living shit out of it, including lifetime warranties and public demonstrations of quality and whatever. If you can't ensure recurring sales by making the product shitty (like most companies do), you need to grow very fast to compensate. Once such a brand reaches high enough trust level, it should not die without making you FU money.
If anyone here is thinking about starting a company making high-quality, lifetime, user-serviceable stuff with well-defined lifecycle from purchace, through repair until eventual recycling / scrapping for parts, please do so; I'll happily put you next to SpaceX and Tesla on the "list of companies I'm an active fanboy of" list.
You could probably find them from my old clients, if they're ever willing to part with the lights. I deal with minerals and mining, now. Maybe I'll get back into LEDs when Cree finally releases their 330+ l/w LED. Till then, there's really no point, LED tech is past HID tech as-is so unless you buy REALLY CHEAP you're pretty much guaranteed to get something that's quality (just watch out for bad power drivers. That's your biggest worry with LED units - they will advertise a 300w unit and it only draws 150 at the wall, and the drivers are under-specced and fail. The LED tech itself is so robust that I've taken cheap stuff like Nichia 219 LEDs or Epistar COBs, hit them with a hammer while they're in operation, pop the protective dome off and flatten the gold bonding wires to the die(s), and it's still lit.