I'd highly recommend against this, from a personal perspective.
It is betting your survival on a dream. The foundation is sand, at best. It sounds like the bank would never give you a business loan so you would have to put a second mortgage on your house. The datasets, or anything produced at all, already has a very questionable commercial viability.
The life rate of 16 lost scientists is questionable in this context. Neural networks are not a new mystery to be solved, and neither is business.