I too ran into this a couple of months ago. Got a panicked call from Purchasing and Production for a specific Bosch Sensortec part. My research confirmed everything you just said.
I dug into it everywhere I could and there was no solid answer. BTW, I got yelled at online by Bosch forum admins for even asking about it.
There was some very confusing info about some of their parts. Example: The Sensortec website's page for a part we use says (paraphrased) "[newer part] recommended for new designs". I've been in this industry for a VERY long time and wording like that is usually an early sign that a part is going to be retired, one step before "Not for Future Designs" and "End of Life". Furthermore, their website tracks their authorized distributor's stock and every single one showed ZERO parts on hand.
However, the suggested replacement is an absolute nightmare - nearly pin compatible but very firmware non-compatible and they even require that you upload an 8KB firmware image on every powerup. (You fetch the 8KB image from github.) 8KB is a sizeable fraction of many MCU's entire nonvolatile memory, WTF are they thinking?!? They have flash on many other devices, if they want to make this thing field-upgradable how about flashing in the current version at the factory and those who care can optionally support in-the-field upgrades?
I raised this question and got smacked pretty hard for "spreading misinformation" (at least I think that's how they described it). I replied with quotes and screenshots of their own website. Finally they openly admitted that the older part is not scheduled for retirement and will continue to be in production. I asked for their projected remaining lifetime for the part and was told "How can we know if it's not scheduled for retirement?" Never mind that most semiconductor companies chart out their product lines at least several years in advance for existing products.
Some weird stuff going on at Bosch Sensortec.
Out of all this I was able to tease out an estimated "next production schedule" of mid-summer. DigiKey and Mouser confirmed they were receiving the same info. Based on that I scoured the grey market and found a reel of 500 pieces and snapped it up to keep us in production for the next couple of months while things continue to sort out.
What follows is 100% theory and conjecture on my part: Bosch provides a LOT of semiconductors to the automotive industry, which has been severely shorted lately due to terrible forecasting at the start of the whole COVID-19 nonsense. They thought auto sales would plummet so they backed off on semi orders, and the fabs (presumably including Bosch) re-scheduled accordingly. Instead, sales have been great and so the pressure is on the fabs to deliver more to the auto industry. Bosch being a huge player in that market, I wonder if (but have zero evidence for) Bosch re-re-scheduled their production back toward automotive, at the expense of non-auto markets. I bet the pressure is intense, and the auto industry consumes a lot of product so they are the 800 pound gorilla that carries a lot of influence over their vendors.
EDIT: Forgot to mention that when we originally designed in this particular Bosch part we confirmed there was a reasonable, pin-compatible second source from STMicro. However, subsequent testing revealed that the STMicro part is utterly unreliable, with failure modes that can only be recovered by cycling power to the chip. We actually added a power circuit to enable this under MCU control but things will have to get pretty dire for us to use the STMicro part... it's just too full of bugs.