The factors that make unionization particularly difficult are first the demographics of developers. The tendency for SW developers to be white, wealthy, and male set a prior political disinclination toward unionization that makes unionization more challenging. Education, tending to be higher, doesn't make up for these other factors.
Secondly, SW developers tend to attract a disproportionate number of self-ascribed libertarian types[1] - the sort who tend to see social relations reified as contracts. As such, a conceptualization of class and class relations aren't a part of their mental model of work.
As I said, I see these tendencies make unionization more challenging, however, there are sectors where demographics and culture may make exceptions. In more exploitative roles, QA and Ops may have greater potential, and sectors like gamedev the ability to commodify workers may ultimately make unionization easier to realize.
Looking toward the long term, projects like Women in STEM, or the myriad others that seek demographic equity in STEM could shift tech population in ways that make unionization a possibility. Sectors that are boarderline now, could find themselves more open to unionization in the future. I'm a bit hesitant to even add this because it's speculative and easy to find counter examples, especially in personal accounts. I'm strictly speaking about aggregate groups at the business org level which is conveniently appropriate to unionization organization and voting. Maybe you dissent, or are in an industry or field which dissents, but a specific counter example to broad trends does not an argument dispel.
1. I'm not placing a truth value of their self-ascription claims and I really don't want to discuss anything more than how this disposition affects their behavior.