I can not understand this either and KnC makes no sense in not offering Saturn/Mercury miners "in mass".
Any ASIC business usually develops a chip and then runs it for at least 1 year. Now that their first design is proven, bugs are worked out and NRE costs recouped, KnC could do large volume runs of the older line and price these things at levels that blow the competition away. Everyone will flock to them since they are proven. Or they could just offer the ASICs and let the 3rd party board developers run with it.
In the ASIC business the high-volume player always wins in the long run, it makes sense for any successful bitcoin ASIC vendor to take this route early to establish themselves. IMHO KnC is opening the door for Black Arrow or other similar ASICs to take a higher volume route with 3rd party board developers.
Doing that may kill the goose that lays the golden egg. I think the reason Knc is doing limiting batches is to avoid driving up the difficulty to the point where it is no longer cost effective for them to manufacture new ASIC devices.
Every machine Knc delivers effectively pushes down the value of their future devices, which is one of the great paradoxes of the zero sum bitcoin mining game.
Agreed that for a given BTC price, there is a hash rate limit that in the long-run limits sales. However, if KnC does not do this someone else will (and are).
You can think of it this way, if the total market for Hashrate is X, KnC could either: 1) fill a large percentage of that total market hashrate or 2) let others sell that hash rate. I'd argue all they are doing by delaying the inevitable is allowing other firms to supply a larger percentage of the market.