Was it that obvious that NVidia would specifically and inevitably be the beneficiary (rather than AMD, Matrox, Evans&Sutherland, or a new entrant)? I mean, it did turn out that way, but I'm not sure that it was obvious in 2014. (Transparently, you can read that I didn't invest in nVidia back then. :) )
AMD blew up, too.
funny you should mention nvidia-- I bought them at $30, because of vr, cryptocurrency and machine learning -- again, it was difficult to explain to people why suddenly video cards were going to be an important thing.
Amazon was losing money because they put every single dollar they earned (and more) into new projects and markets that expanded their reach.
Definitely agree with your second paragaph though! There are so many companies that for some time fly under the radar of the general public, where experts in the field can already see their success coming a mile away.
Nvidia would be another example. Anyone in Computer Science around 2014 could have predicted their meteoric rise due to machine learning and CUDA. But people outside of the field would have no idea.