So, if Ricoh bought Tamron, why would they come up wit a different answer - that somehow it would now be profitable?
As a consumer, I get why Pentaxians would like this - but from a business perspective, it makes zero sense.
Sadly, I tend to concur.
The caveat would be that, if a lens stream augmented by Tamron would increase Pentax body sales, then it really doesn't need to be profitable in itself, just break-even. That would still be a win overall.
Against that synergy, however, is the other stark reality: while a lens-maker can save money by offering the same essential glass design for multiple mounts, there isn't much (any) overlap between DSLR lens designs and ML lens designs--the registration parameter differences are just too large. So design effort achieves no synergy regardless of who does the designing.
Passing or sharing DSLR design expertise to/from/with Tamron, then, isn't going to make designing DSLR lenses any cheaper. While we commonly say that Tamron and Sigma stopped supporting Pentax, the harsher reality is that they stopped supporting DSLR design entirely.
yes, that's certainly part of it. A big part of manufacturing is maximizing your machines/processes. It's why Tamron and others abandoned Pentax lenses to begin with - each mount adds cost in R&D.
When running a line, I presume there's a change-over cost to switch from one mount to another as well.
When running manufacturing you want to get as close to your max capacity. In simple terms, if you have one process station to make a lens (not reality, but for a simple concept it helps) and that station has capacity of 1,000 units per day without a change-over and a change-over results in a reduction of 400 units.
So, if you have the demand, you definitely want to produce the 1000 per day. You want to do minimum switchovers, so you don't incur that 400 unit penalty. You certainly don't want to do the switchover to Pentax if you only have demand for 300 units if you can keep producing Nikon lenses and make/sell 1000 of them. And this is simply the manufacturing process cost - not materials, R&D etc.
Eventually, if you have enough demand you start increasing capacity - but that's capital investment in new equipment usually. So, there has to be a pretty big pay back.
As you mention, part of the benefit to the Tamron, Sigma model is - a good portion of the materials, R&D and manufacturing process for a lens is common across all mounts. So, it's not as expensive to offer multiple mounts. But, the scenario I described above still comes into play for part of the process - building the mounts and presumably attaching them. Even before mirrorless, Sigma, Tamron etc stopped producing K-mount because they had enough demand from other mounts such that they didn't have unused capacity in their manufacturing stations. Or not enough unused to offset the added cost of adding K-mount to the equation.
If Ricoh was going to buy that manufacturing capacity, it makes zero business sense to under-utilize it to focus on K-mount. Pentax buys wouldn't make up for the lost Canon, Nikon, Sony sales.
Now, as you say, the problem is even worse for K-mount. Now, you lose the design, parts and manufacturing synergies because the other 90% of the lens is different to. But, it was this way before mirrorless because the demand for new K-mount lenses just wasn't high enough. It's a tough nut for Ricoh to be sure. They need new lenses but the core of their user base loves buying used and continuing to use 20 year old lenses.