Is this a globally significant wasted funding? If not, it sounds like a reasonable idea to try something new from time to time. Worst case, some project trial fails and we learn a new failure mode / limitation. Best case, we get a new reliable power source.