Disclaimer: I am not and expert and not even an amateur in the area, just an interested layperson. Also, consider reading everything Scott Aaronson writes on his blog about this topic.
Basically, there are several known tasks on which quantum computers perform better than classical computers, such as simulations of quantum systems (like that quantum chemistry you mention), quantum-certified randomness, some types of authentication, encryption and secure transmission, factoring, database search and a few others. They are no better than classical computers at general NP-complete tasks.
So, odds are that there will be "quantum modules" incorporated into the usual classical computers (and by classical computers I mean the successors of what we now think of as smartphones, e.g. iPhone 30q or qPixel 2040 or Galaxy Q50). The user-visible effects might be something like a "quantum-secure padlock" in your browser. On the server side there might be something like quantum search modules used by the likes of Amazon, Google and Oracle.
I'd also bet that there will be at least one or two unexpected applications of QC discovered in the next couple of decades, judging by the progress in the field, but it's hard to guess whether they will have any revolutionary impact, rather than an incremental change in a specific area.
The database search thing is, according to my understanding, widely misinterpreted. As Wikipedia says:
Although the purpose of Grover's algorithm is usually described as "searching a database", it may be more accurate to describe it as "inverting a function". In fact since the oracle for an unstructured database requires at least linear complexity, the algorithm cannot be used for actual databases.
To actually build Quantum Postgres, you need something that can store an enormous number of qubits, like a hard drive.
Given the cooling requirements, I would expect cloud computing to be done as a service, similar to AWS. You can shrink a circuit, but not a refrigerator.
The first application will probably be to break most current cryptography. I'm sure there are plenty of governments that have intercepted communications and are holding onto the data until they can decrypt it. Also bitcoin mining and generating hash collisions.
Beyond that, the main applications I can think of are solving optimization problems and quantum simulations. The optimization applications are to broad for me to speculate about, but the simulations will probably be useful for protein modelling and design. The other main application I can think of is for materials science. Modelling the properties of materials is currently pretty hard and heavily determined by quantum phenomena.(Solids wouldn't really be a thing if electrons and nuclei obeyed classical electrodynamics.)
I wouldn't expect the average consumer to have a direct need for quantum computing. At least, I can't think of anything off of the top of my head. Rather they'll see the downstream effects of QC. New and better materials, new encryption algorithms, better medical technology.
Quantum computers aren't going to be owned by individuals for a long time, because of the requirement of cryogenic cooling. Instead, you'll rent time on a quantum computer, or use a service that itself uses a quantum computer.
Anyhow, what might be the equivalent of looking at ENIAC and predicting flappy bird? Audio processing software that uses fast fourier transforms, maybe? Least-action optimization algorithms that power the AI that recommends what videos you should watch?
It's hard; quantum computation is just more specialized (in our high-temperature environment) than classical computation.
In the early days of electronic computers, machines like ENIAC were used for niche applications such as calculating ballistic trajectories and simulating nuclear explosions. These applications are very far removed from what computers are predominantly used for today.
It seems we have reached a similar development stage with regard to quantum computing. The applications researchers cite, cryptographic analysis and quantum systems simulation, are once again very far removed from everyday life.
Does anyone have a prediction on what quantum computers will be used for once they become affordable enough for regular people? Or will it forever remain just a research tool? Are quantum computers the key to cracking quantum chemistry and thereby molecular nanotechnology? (this is my guess as to what the actual impact of quantum computing will be)