Back in August 2017 when Purism started, the only way to get a smartphone that ran on 100% free software was to either use the ancient i.MX 6, which have given horrible performance, or to use the upcoming i.MX 8M Quad which has taken years to get decent mainline Linux support (it still doesn’t have suspend to RAM) and critical things like its MIPI CSI2 camera interface are poorly documented. The other options were to use the A64 like the PinePhone did and hope that Mesa would eventually support it (which ended up happening in April 2019) or to give up on hardware kill switches, publishing open source schematics, and 100% free software, and use an older Snapdragon that has mainstream Linux support.
Given the choices, I think Purism took the only path that was truly capable of generating a large number of pre-orders, which was necessary to finance its dev work on Phosh. Without 100% free software and hardware kill switches, I don’t think Purism would have gotten enough pre-orders to pay its software developers, and it would have to use Ubuntu Touch/Lomiri or Plasma Mobile, which both have major problems in my opinion. Lomiri simply isn’t being well maintained–it only had 112 commits in the last year and only has 9 people with commit access, so basically Purism would have had to take over maintenance of the project which would have been very expensive. In the case of Plasma Mobile, it wasn’t at all complete in Aug. 2017, and it would have required a lot of dev work to get it ready, which also would have been expensive.
In other words, there were no good choices for Purism, and I frankly think that Purism chose the best course of all the difficult options it had. You might want to read my essay on why Purism created Phosh.