Android phones haven't just started using the SGX540, the Galaxy S used it and it launched shortly after the iPhone 4 in the United States.
I must have forgotten about that, my apologies!

I was recalling reading a review for a recent phone that had the 540, and I didn't remember anything else with one. Looking up the Samsung Hummingbird SoC confirmed it.
... And it's still not in a phone!
You don't have to tell me that... I'd like to replace my phone sometime soon.

Of course, that's without having to repurchase all of the applications that I've already spent money on.
However, from what I've read the PowerVR SGX 543MP2 from the iPad 2 won't be used in the next iPhone; the single core version (PowerVR SGX 543 I believe) will be used due to heat and battery life.
Hmm that would be a bit of a bummer but understandable. The 543 (single core) has 2x the geometric processing of a 540, which makes it formidable. The GL Benchmark 2.3 tests that I saw on Anandtech (I forget which review it was) showed the MP2 in the lead using off-screen rendering, but I doubt it will keep that lead without the dual-core variant.
Although, you bringing that up makes me wonder... if they can't handle a dual-core variant... how will Sony manage the quad-core in the PS Vita? The Vita is a little bit bigger overall than most phones, and the battery life information from the Tokyo Game Show certainly came as a huge bombshell (~3 hours of gaming). Although, I'd be surprised if I got 3 hours of gaming out of my cell phone. I don't think I could even play Kenken for 3 hours straight without pretty much killing my phone, and that's a 2D Sudoku-like game!
It has the fastest GPU. Its CPU is actually below the current dual-core average (Tegra) in tests I have seen.
Yeah. I could have sworn that I put "GPU" into that post, but I must have intended to and forgotten. Given the very small differences among the A9-based CPUs, the performance mostly comes down to raw speed, and 1Ghz isn't going to win any contests there.
Comparing SoCs is a hard thing to do in this case as well. Usually GPUs on review sites are compared using as similar of a test bench as possible, but given how some GPUs are only found on some SoCs which are only used in conjunction with some mobile operating systems... you get a pretty big mess. A good example is the 543MP2 in the A5. The only product that it can be compared to is TI's OMAP 4470, which I don't think is even in anything yet. It's technically the 544MP2 in the 4470, but they're pretty similar.
But anyway... maybe with that, we can finally get some Apples to Apples (or rather... Android to Android

) testing.