Mentioned to Sath last week and thought I'd share publicly, as since that conversation it's increased by 44%.
Golem (GNT) is an ERC-20 token (aka, Ethereum-based) that I believe has a lot of potential. They're looking to create a worldwide decentralized network of computers, which any individual/group/company can contract for processing power. Likewise, any individual/group could configure their idle processing power to automatically sell on the network. It's like folding@home, but for any CPU/GPU tasks & with profits for contributing.
It started off very pie-in-the-sky-esque with nothing but the idea, a basic proof of concept, and very slow progress beyond that... but they're finally starting to deliver. A working alpha was released in the Fall, which shot the price up by 400%.
I've been following them for the better part of 8 months, and if they pull this off it'll easily be another big-ticket coin.
https://golem.network/
https://www.reddit.com/r/GolemProject/
At the very least, keep an eye out for the Brass Golem 1.0 release date for Q1/Q2 2018, as that'll be the first working version where people can earn real currency by selling their idle processing power (though only for certain tasks). If it's a success, the coin will no-doubt shoot up in value.
Worth noting that this used to be a top-10 coin, but it fell off the radar (due to the slow progress and no working product, as said above). If they jumped back into the top-10 right now, it'd represent an increase of over 1000% from their current price & market value.
Tough to say whether they can pull that off of the immediate future, as the the 'final version' is still 3-4 years out, but if they succeed in rolling out Brass Golem and keep making progress it's definitely possible. I can easily imagine this becoming a valuable alternative vs. the massive costs it takes to build, run, power, cool, secure, update and maintain your own super computer/data centers. Like, the amount of money my company spends on our super computer is ridiculous, and it hasn't stopped running since day-1 and already has an insurmountable backlog... and it's basically like this across the research industry.