I remember back when Nvidia was aiming to aquire ARM, there was a whole thread of discourse that I think even Huang himself had spoken to, about how the acquisition would transform Nvidia, would remake who they are and what they do. ARM themselves have rarely been an exemplary citizen, with driver support for their GPUs for example having only gotten respectable in the past couple years. But ARM was still a multi-lareral company. It would have been a different NV: Nvidia as head of an ecosystem, a steward for many, versus what Nvidia has been, Nvidia as a moat keeper with only their bespoke Linux4Tegra (L4T) for example. That was going to be the big change.

Qualcomm certainly seems to be saying that they want to be a different company. That they want market-share among people building products. The example elsewhere of the e-scooters using RPi's seems like the market space Qualcomm is striving to open up.

Your middle paragraphs capture a lot of the sentiment. Qualcomm is a hard company to trust. There have been a lot of neat weird interesting things that have gotten mainlined, and it's cool to see, but most products are incredibly hard to develop for, push you into vendor Board Support Packages, and don't have docs available. This chip similarly lacks technical docs.

But it sure is exciting to think maybe Qualcomm might actually want embedded market share beyond the high end of phones, routers, and laptops. And if they do want this market share, they're going to have to change.