From what I have seen, there are quite a few kits that can be either a plasma cutter or router. I think it would be tough to take the time to build one and try and run it to make a profit. Most of the ones I have seen, the person has done it because they want to, as in that is the goal, not to use the machine to make money.
I went though lots of different scenarios before I bought mine, one was to buy two old Shodas that had been in storage for 15 years. I decided to pass even as cheap as they were because I needed to have a running machine on the floor, not a project. I think an 80/20 kit with a hand held router motor would possibly be better than nothing, but barely. I honestly don't know how tight of a tolerance they can hold. I envision him getting it together and then realizing that he still has to go and redo and tweak the parts to put a box together. He will also have to deal with not having a toolchanger or an option to have other tools without changing them mid program. Even with a tool touch off it would suck. I think if he is going the kit route, I don't see it being any faster or better quality than using a slider or panel saw and drill tub, because he is still going to have to do lots of 2nd ops manually and possibly resquare the parts etc. At that point, I don't see a reason to do it. Now if he planned on doing mostly curved or artsy stuff that wouldn't be as big of a concern.
I bought a '92 Komo VR805 multispindle machine. It was $19k from a dealer where I got to go test drive it. I had alot of experience running one just like it, had the post processor etc all ready to go as well, so it was painless to start making money with it. It has 8 spindles that I can use like a toolchanger or I can use it like a duplicator. I do cut cabinet parts on it as that is what I have, but to be honest if that was most of my work, it is not the best machine for it. It doesn't have a drill bank on it, so when I do a closet side with a hundred plus holes in it, it takes forever and a day one at a time compared to a machine that could do them 3,5,7,9 at a time. I honestly spent more on it than I should have, but I couldn't gamble on one that was unknown, I needed it going as soon as possible and the one I got I had making money within 2 hours of setting on the floor. There are the exact same machines on ebay right now for less than that, but they may not be powered up etc, so you roll the dice.
Back to the hobby machines, yes he could probably make cabinets with them, but there are so many corners cut on them that it will be harder to tell realistically if it is profitable. The tormach example is a good one, they work, but not at a level that inspires confidence. The same issues that hamper a tormach hamper a router. If it is not rigid, you give up feed rate and DOC, then you put heat into the cutter and you give up tool life and sometimes actually burn the material, which can catch on fire... If it happens to be rigid, but not enough hp to run a tool properly, same results.
Now the feature that most people overlook. If you can't hold the part, you can't cut it. Most of the cheap machines I have seen don't have a vac table, they use a few t slots and you have to clamp the material down to the table, then move them around so you don't cut them and all that fun BS. Those tables typically are thin extrusions that aren't the most rigid either, so you can get deflection in Z
.
The next step up machines that do have a vac table don't have a very good one and you have to be very cautious and very methodical and maybe do a song and dance and sacrifice something to the vacuum gods to keep the parts held. Some of the reasons are the pump itself isn't big enough, good ones aren't cheap. And anyone that believes a regenerative blower is a vac pump needs to their head examined. The cheap machines tend to have one big zone instead of multiples. By that I mean you can't section off zones on the table to apply vacuum to certain areas. If you are cutting a full sheet all the time then not a problem, but if you want to run half sheets or smaller, then it makes life much easier if you can turn on the specific zone you need.
Then the physical table itself, many aren't made with any thought to airflow, so the vacuum pump may be good, but it is hamstrung by the table design and the plumbing of hoses or tubes. The better the flow, the better the vacuum will be applied to the parts. The design of the actual table top makes a difference as well. If it is just a open plenum with a few islands holding up a pc of mdf, it will be much harder to do less than full sheets. You'll have to to tab most of the parts, may have problems holding anything besides melamine, especially if it is warped or bowed any. The better ones have some version of a phenolic top that has a grid cut into it and vac holes. It sets on the open plenum and is usually sectioned off for the zones. The grid is typically larger slots around larger islands. Then the one I like best is an aluminum top with a grid of 1" square islands and grooves 1/4" wide and 1/4" deep. The vac holes are app every 6" and the islands have a shallow cross cut in the top to allow more area for vac to reach. This is really nice because you can use oring material and section off a small area or weird shape etc to apply vac to only where you want it. It is pretty fast to changeover. You can do the same thing with dedicated fixtures, but for low volume the oring grid is cheaper and faster. (See NEMI grid tables)
I think instead of spending $10k and and unknown amount of time building a kit, he would be better off to get a good used machine for $15-25k.
One thing he could look into is to sub out some parts to see how they are processed and if he likes the results etc, then based on what he sees them do, he can pick what he wants a little easier.
Thought of something else, a 5 ft vs 4 ft table is nice if you ever want to cut any real baltic bich ply in the 5x5 sheets, you can also get 5x_ sheets in most anything else too and it can really help the yield on certain jobs.