It's all a matter of timing.
We know at some point petroleum will become impractical to support the energy needs of civilization as we know it. In 1979, people thought the end was nigh, when in fact it was just an oil cartel flexing its muscle. But the ability of the cartel to do this was a harbinger of peak oil. The US was no longer anywhere near energy self-sufficient. The same thing happened in 2008, suddenly the end was nigh. It wasn't.
Let me draw a closer analogy. During the Dot Com boom, lots of individuals made money, but very, very few web enterprises did. VCs were pouring money into startups with nothing more than a name and a PowerPoint presentation -- not a real business plan. There wasn't time. Money had to be spent. Why? Because everyone knew web commerce was coming. And they were right, web commerce was coming. Yet the vast majority of money invested in those days was wasted, at least from the perspective that investments should provide returns to the investors. In part it was classic bubble behavior, but even *solid* business plans had a huge element of uncertainty. If the business plan was timed *right*, it would be like buying Microsoft in 1986.
Some day, if the conditions are right, one of these photovoltaic "breakthroughs" may take the world by storm. The chance of any single "breakthrough" doing it is vanishingly small, but the chance of *some* "breakthrough" doing it is substantial. It has to be the right development at the right time.
There are basically three avenues I can imagine working:
(1) Really cheap cells. If it works and can be produced commercially, this is an almost certain winner, because there are many places we can use power and are throwing away photons. Probably won't change the world if it is really inefficient, because *other* costs around installation will limit success.
(2) Incrementally better cells. If it works with given production infrastructure and doesn't have any significant drawbacks, a nearly certain winner that will expand the utility of photovoltaics to applications that are currently marginal. Certainly not a big winner in the short term, but a step forward for the tortoise in the race.
(3) Radically more efficient cells. Least likely to succeed, IMHO. The best cells on the market are about 23% efficient; given thermodynamic limits there isn't a *great* deal of room for improvement. Super-efficient cells will probably have niche applications, but any "breakthrough" in this area needs to have the proviso "and costs about the same to produce as current photovoltaics" before it's worth sitting up and taking notice of.