one of the unanticipated costs for me was marketing, I sell my webbased software to businesses and its not the kind of thing you can sell via search results & adwords campaign, so alot of brochures, trade shows and cold calling are required to sell subscriptions.
I know nothing about electrical engineering but there was a documentary on the history channel about this and they interviewed the leading researcher(a professor in the US) and he basically said no ones ever built a tower the size of Teslas to actually test if his theory would work, it didn't get into the details of why the experiment couldn't just be scaled down. But I think he implied there was some sort of critical size you need to get to where it would become efficient.