I'm not sure how viable this is. Linux phones already opt for hardware that's as open as possible, i.e. they use parts with the most open documentation and drives, but the trade-off to that is that those parts are functionally already end-of-life when they're in the phone, either because it's an old design that's been opened up to squeeze a bit more money out of an old design, or the design was third-rate to begin with. Not to mention that the baseband side of things is closed no matter what, so the phone that's completely true to the FOSS ideals seems impossible to make no matter what. And who would buy a phone with a third-rate chip and battery life? And since very few people buy them, prices aren't able to drop any significant amount.
I understand why people aren't willing to make a devils bargain in order to make a decent phone first, and then put Linux on it second, but I can't see any other way for this to happen, other than the phone market magically becoming more open somehow. If you could install Linux on any phone, since all the drivers are already out there, then we wouldn't be in this pickle, but every single Android phone out there has a different set of drivers and very few of them are open and possible to implement without an enormous amount of work, unlike the PC world, were at this point, only the really weird stuff (and Wifi from certain vendors) doesn't have some form of Linux driver.
Separate from baseband, the (sub)device closed firmware blobs are non-ideal, and eventually you'd want open source even for those, but maybe don't have to be a high priority. Mainlined open source for corresponding drivers are much higher priority. Even Debian now tolerates such blobs.)
Together, these facts make it so that (competitive) wireless modems require organized businesses to create, and organized businesses don't want to share their code with competitors. A foundation dedicated to creating open hardware and software for a competitive wireless modem would face giant hurdles both in regulatory terms, and in hiring people who can actually work on this extremely difficult technical challenge.
Also, building an open source software for controlling wireless modems that complies with the law is probably not fully possible. Per law, to sell a wireless device, you as the manufacturer are responsible for taking reasonable precautions against users misusing it to emit in reserved bands, or to not respect military device priority in the allowed bands. If every user is extended the rights and documentation for modifying the software as they see fit, you're clearly not taking reasonable precautions to prevent them from breaking the law.