Infinite, from your perspective of being able to consume it. The abundance is sufficient that from your perspective - and even to that of the species for a thousand generations once starting on the Oort - to be considered infinite. Ofc nothing is truely infinite. Stupidity comes pretty close.
With regards to rate of return, it should be incredibly easy to ensure high rates. A few tonnes of mining equipment can be expected to provide megatonnes of output between failures. The one-off cost is comparitively steep to buying similar resources on Earth for the first haul it throws back but each subsequent is continually reducing this cost. Work out the cost of getting 10 tonnes to a transfer orbit, then how much xenon(assuming EM-Drive unsuitable) you'd require to move 10 tonnes centrafugally accelerated from just past Mars back this way. It's a lot cheaper to launch the xenon from the floor and use that to get mass back here, especially a few thousand tonnes in. It will not take many subsequent hauls to reach a level of attribution unfeasible with lifting mass otherwise. Sale of say 60% of returned mass would easily pay for more xenon and other consumables. Especially if we pick rocks with "high value" elements like platinum in abundance. Once these things are cloning themselves then the rate of return becomes exponential, and incredibly difficult to compete with.
I might of stopped to consider the technologies involved with making this plausible - which could be how I've considered this is entirely plausible. I might even be playing with some of these technologies.
Clearly, these "tugs" will require the ability to self navigate. Manual piloting is inefficient and due to telemetric lag unfeasible - when a few thousand units in, also impractical. This does not represent any form of a problem, the technologies involved in navigation have been in use for decades. The "AI" involved would require to be barely smarter than an ant, and behave remarkably similarly. Repairing themselves is a tall order IMHO, far more feasible to recieve maintainence when it drops off payloads.
Extraction of ore is also a pretty trivial matter, rendering it into a powder should be easy via a combination of application of resonance and mechanical grinding, as is sorting and grading/purification. centrafugal sorting should be more than adequate to sort by atomic mass for individual elements, and then again for individual isotopes.
The reason behind towing the rock is that after placing facilities in orbit, cloning these adding expansions and throwing past mars, the equipment in situ Earthside is still usable. And those rocks I mention need pulling out of problematic orbits. Much of them come close enough to not bother chasing them with the mining facilites, and instead park them next to the facilities. This is ofc a lot more feasible if able to entertain the EM-Drive, as that can continually output and given enough time parking will be assured. This isn't to obtain the resources in these rocks, that's a bonus, it's just to get them out of the way. With regards to the asteroid field and mining of, it's likely more suitable to move the facilities to the general location and tow individual rocks to.
The endavour is still far far cheaper than considering lifting a few hundred megatonnes from the surface. The recyclable materials in LEO are more than sufficient. The few thousand tonnes of titanium and aluminium from upper rocket stages alone can build the chassis and panelwork for a hundred thousand "tugs" on top of the few hundred seed factories potential of which I intended one. With reference to the size of the rocks, you don't start with the biggest ones for harvest. The harvesting of the smaller ones resulting in material to build for larger scales. Rather than scale up the tug, just deploy more of them and have them work collaboratively. Or make the rock smaller.
You might not be able to see a way this can work - but this doesn't detract from the fact it will work. Developments required to existing technologies are incredibly minimal, and where required achievable. To actually take into account the return, the ROI is impossible to be matched. Especially as the initial investment is to be met by itself, and therefore 0. The costs involved in the development I'm more than prepared to shoulder - and will be anyway.
Yes, fleets will be required, and these get built then they exist. A few, or parts for, put up with the seed factory can tow in scrap to build more tugs. I'm fairly confident that this has already been adequately covered previously.
Installation of electrical systems doesn't represent any issues, building them could be as simple as printing a board, pasting it, splatting components into the paste with a pick n place machine and taking a quick trip through a reflow oven. Actually making the electronic components - ie: chips - is something of an issue, not so much as an obsticle as a hinderence and likely represent some of the later to arrive functionality so therefore be the bulk of "consumable lift" required after the initial lift (unless forced to use propellant, then it'll likely 50/50 fuel/parts). It is not impossible to provide for this ability, expecially once there's a steady flow of sillicon to melt up and grow into crystals, but until there's this supply then it's likely useless having this capcity - and as the larger mass for this addition can also be constructed in orbit (whilst waiting resources to return from past Mars) it can cut down initial lifting cost.