In reality, they can't do this until they have a working quantum computer of an interesting size.
For now, it's limited to toy-size QCs, which can be simulated faster and cheaper using an ordinary computer. So if they are on a cloud server, it's more for curiosity value only. They aren't actually useful.
There's a lot of buzz and hope and money going into it, but nobody knows with certainty if it's even possible to build one of a useful size.
The "quantum supremacy" claim which happened already was genuine and deserves accolades, but it turns out only to be useful on a very contrived kind of problem. The clever part, in a way, was figuring out a suitably contrived problem, to show that quantum supremacy is technically possible without resulting in a useful quantum computer, and it doesn't run any of the standard quantum circuits.
There are different implementations of "quantum computing" which are appropriate for different types of problems. Google and Wikipedia will teach you about them, for example quantum annealing vs. universal quantum computing.
As for what types of problems: D-Wave's quantum computers (and adiabatic quantum computers in general) are best suited for optimization problems. Some examples are scheduling, graph coloring, and other NP-hard problems.
AFAIK quantum computing, as a whole, is very far from breaking any kind of interesting crypto.
Just thought I'd mention, quantum computer design started in the early 1980s, and many of the theoretical results came through quite a while ago.
What's happening now is the investment and funding landscape has changed considerably.
I'm sure that once the electrical engineering side of things is in place and commercial machines are being manufactured (custom or model based design), it will become a new consulting domain for software engineers to work in.
What types of problems are quantum computers anticipated to solve? I've only heard about how they are able to break certain cryptographical algorithms that were designed long before quantum was even a thought.