I suspect if Sony had an ultra high resolution sensor ready to launch in the near future we would have seen it in this camera where it could potentially make maximum impact on the market.
I'm guessing Nikon will be quite happy that Sony hasn't really been able to offer a headline feature to best the D850 here or undercut them in price significantly.
That's exactly why they've launched these products when they did.
Part of it is defensive - limiting the impact of a rival's release. Part of it is offensive - expanding market share. And part of it is profit-grabbing - maximising sales. All three aspects have to be done in balance.
Sony has all the necessary components and designs available to make any of a number of cameras - fast/low-resolution, medium speed/medium resolution, slow/high-resolution, all with similar features in other respects (AF, construction, etc.). It's just a matter of finalising and releasing each one at the optimal time. But they can only release each one once. Each camera design is like a round of ammunition in Sony's arsenal - an expendable resource that must be saved for the most opportune moment. If they released everything at once, it would likely be of limited effect, and leave their magazines empty until R&D can replenish them.
It made sense to release the A9 first. It came out a year after the D5 and 1Dx2, and well before the next generation is due. Sony couldn't have released the A9 at the same time as the D5/1Dx2, since it wouldn't have been anywhere near ready. So the A9's release was a pure offensive move, designed to win market share and earn Sony a foothold in the fast action market. Prior to its release, Sony didn't have a demonstrated capability to make a fast-focusing, fast-tracking camera suitable for shooting sports. No-one knew whether the A9 would be able to match the AF of Canon/Nikon sports cameras, or existed more at the level of the 5D3/5D4 general-purpose cameras. It had to demonstrate that capability to the public somehow. Any of the possible combinations - fast/low-resolution, medium/medium resolution or slow/high-resolution - could have demonstrated this, but the fast/low-resolution body made the most sense to release. For one, it had the lowest opportunity cost - the A9 doesn't really compete with the A7r2 for user base, since they serve almost opposite ends of the photographer spectrum, so self-cannibalisation would have been less of an issue. Secondly, it expanded the Sony lineup into, and made Sony a viable (if niche at this point, given the lens selection) competitor in a new area of photography, where they had previous lacked any presence. Thirdly, it made much more sense to release the fast/low-resolution body earlier in the product cycle of the 1Dx2/D5, when neither Nikon nor Canon would be ready with an update that would thoroughly trounce the A9. And, finally, from a profit-grabbing point of view, it made more sense to release key features in dribs and drabs rather than all at once. In this case, fast AF was the key feature (dual slots, better mount and bigger battery are all nice, but would probably not have sold a new body by themselves without the AF). It was not available in the A7r2 - in order to get better AF, users would have to buy the A9. But the A9 doesn't offer 42MP (20fps is nice to have, but it's a relatively niche feature that, for most applications, isn't as generally useful as 10fps/42MP). For many buyers, 24MP is enough, though. By doing what they did, they forced anyone who didn't want to wait for a fast-focusing Sony to buy the A9, while still leaving room for repeat sales in the future from either a fast-focusing, high-resolution model (with users buying the A9 first for the AF, then selling it and buying the A7r3 or A9r when it came out and was more suited to their needs) or a budget model with the same AF but slower frame rate (A7Mk3, possibly). This wouldn't have been the case if they had released one of the higher-resolution bodies first, nor would it have been the case if they had released an A7Mk3 first.
I suspect that Sony's recent release of the A7r3 was a defensive move, to limit the impact of the D850 by introducing an equally-capable competitor. Whether it was planned way ahead, or only in the last few months, likely depends on how much Sony Camera knew about the capabilities of the D850 - they may have known something about the sensor, but possibly not the rest of the camera. Certainly, if the D850 had not been released or was much less capable, I would have expected the A7Mk3 (24MP like the A9, but slower-shooting and possibly with weaker AF) to come out before the A7r3 (basically a medium-speed, medium-resolution version of the A9). But a 24MP/7fps 'budget' camera would have left the D850 go unchecked and allowed Nikon to run riot amongst the biggest category of high-spending user (those who need some speed, but not necessarily 20fps, but would also like some resolution to go with it), with no Sony body able to match both the resolution and the AF of the Nikon (let alone the frame rate). In fact, it was rumoured until very recently that the budget model would be what Sony would release next - perhaps it was a last-minute change of plans (presuming that any of the designs are ready to go the moment management decide it's time, with a few months needed to switch around production lines, etc.).
Most likely Sony hasn't released the high-resolution/slow-shooting version yet not because it doesn't have one, but because it doesn't need to, and because more profit can be made by releasing it in six months' time. They could have released it instead of the A7r3, but, not only would it have been less profitable to do so (a significant number of people will buy a 42MP/10fps body now as well as a 70MP/5fps body later, either selling or keeping the older body; it is likely that fewer would do the reverse, assuming AF capability is the same) but a 70MP/5fps body also wouldn't provide the same defensive barrier against the D850. It would have left Sony with a fast-shooting, low resolution action body and a slow-shooting, high-resolution body, but left the entire middle ground to Nikon.
I suspect we will see the high-resolution/slow-shooting body in 6-9 months time, after all the early-adopter/impatient buyer A7r3 sales have been made - just in time to meet Canon's 5Ds2 (possibly pre-empting it, and, if Canon hasn't stepped up with the capabilities, rendering it a flop).
And no doubt Sony is working on the next generation already, to meet the Canon/Nikon mirrorless challenge. The current generation - A9, A7r3, whatever the high-resolution/slow-shooting one is called and the A7Mk3 - are all ready, to be put into action when the time is right. R&D is almost certainly focusing on the next generation now, in time for the Tokyo Olympics.
