Interesting sidenote: I saw a reply to this post on Twitter where SourceDNA commented how the makeup of the top 100 apps differs from the long tail of apps:
Developer behavior is pretty varied along a number of angles. Since even indie apps can hit the top ranks (especially in games), it's a little less useful to talk about apps by rank and instead more by company background.
Swift is growing but not huge yet. Certainly there has been a flood of experiment apps in the long tail, where someone used it as a learning exercise. Some of the larger companies are using it for a new project once in a while, but it's certainly been most popular in games.
What never ceases to surprise me is the influence a single developer can have, even in a large organization. Someone picks up a tool and then suddenly it's used everywhere in the company.
Lyft was completely rewritten in Swift so that was news. Many other companies could be using Swift but it's not really news if 95% of your code is still in ObjC.
One thing to note: Carthage usage is invisible, since frameworks built with Carthage are just Plain Old Frameworks. So people that have migrated from Pods to Carthage will be hard to track.
Similar to the cocoapods note, my guess is it's catching on much faster with the long tail than the most popular apps.