When Element X first launched, the goal was for it to cover the personal messenger use case. This worked just fine for some people, but for many, feature parity with the old apps and parity with the web experience was a hard blocker.
Both Spaces and Threads are about to land, and there are other lower-profile features that also need rounding out. We would expect full parity by April this year. At which point migration should be an obvious choice.
The complexity associated with Matrix comes from wanting to build a multivendor ecosystem around an open standard, the decentralisation (and federation) to avoid building islands and then implementing things like multi-device e2ee and VoIP in such a context.
However, it is exactly these properties that make it so appealing to an organisation like the EC as they pursue digital sovereignty.
The goals of Matrix have nothing to do with being a social network. You could theoretically build a social network on top of Matrix (Matrix essentially syncs JSON in real-time), but I'm not aware of a project with traction, and more to the point, those projects are not relevant to this discussion. Yes Matrix is resilient to network partitions (that's a good thing for a messenger!), but that seems entirely orthogonal to your point on social networks.
Finally, I don't think it's fair to draw a comparison between the relationship between Kandra Labs and Zulip vs Element and Matrix. Yes Element is a major player in the Matrix eco-system having originally formed to hire the Matrix founding team, but since then many Matrix vendors have sprung up and if Element were to disappear tomorrow Matrix would continue. In fact, this is the whole purpose of having an independent Foundation, which in turn encourages multiple vendors to operate side by side.
Both Spaces and Threads are about to land, and there are other lower-profile features that also need rounding out. We would expect full parity by April this year. At which point migration should be an obvious choice.