Arguably they could have avoided Microsoft "betrayal" as well. Ultimately, Windows 3.0 was a skunkworks project of single engineer against the corporate decisions of Microsoft, and only when it was quite closer to complete did it start getting management buy-in.
Pretty sure for a long time it was "cloaked" as stop-gap solution, a continuation of the lesser-known "windows runtime embedded in application" option that some software shipped with.
They had different goals. But it's not clear that Microsoft's goals--a largely hardware-independent OS--ever made sense for IBM. As it turned out, a more proprietary PC architecture didn't really make sense for IBM either but that was sort of beside the point.
Pretty sure for a long time it was "cloaked" as stop-gap solution, a continuation of the lesser-known "windows runtime embedded in application" option that some software shipped with.