And as xorcist said https://news.ycombinator.com/item?id=22287837 it seems likely that Nokia really had finally figured it out with Meego (though I've never touched the thing myself). It was fairly inevitable that a decent #2 smartphone OS would be along sooner or later as soon as iOS had provided the competitive impetus and the blindingly obvious model to crib from.
OTOH it's also quite clear why Google would not have wanted to gamble on waiting around for somone else to get their act together, especially since it already happened to own the guts of an okay smartphone platform. Not only would a viable #2 have taken an uncertain amount of time to emerge, it could have turned out to be someone quite unfriendly to Google: it would very likely have been MS, after all. And after all (viz. __ka https://news.ycombinator.com/item?id=22287118 ) what other platform owner was ever going to be friendlier to Google search than Google itself?
And let's be frank, by the mid-2000s it wasn't really that bloody hard, from a purely technical point of view, to ship a decent mobile OS running on an adequate hardware platform. It seems pretty clear that in almost all cases, what was most sorely missing was someone with both the power and the taste to make an overall assessment of UX and say "no, this is shit, come back with something acceptable"; maybe to even make a few hirings or acquisitions if necessary to get someone competent on the problem. But this is one facet of one of the most remarkable and slightly strange things in the history of tech. Even by not too long into the '90s it was easy, for anyone with eyes to see, to look at a Macintosh on one side and a VCR blinking 00:00 on the other and see what the future of device UI was. But companies with massive resources and their very future at stake, including firms of impressive competence like Sony and Nokia, couldn't get there. Up to the early iPhone age it seems that the list of companies which could do good work in software-heavy/"smart" UI was roughly:
1) Apple
2) Microsoft, but only when Apple first provided them a detailed model to crib from. Not something to sneer at too much, because others couldn't do it even then
3) Some Mac/PC ISVs who could do good work, but purely in software and only inside the lines of the existing Windows/Mac platforms. (OFC by this time MS itself was effectively a Mac ISV with notions.)
4) Some device or HW startups who flailed around commercially (Danger, NeXT, Be, the set-top-box group at Sun which begat Java http://www.blinkenlights.com/classiccmp/javaorigin.html ...)
See also the iPod, Apple's Digital Hub strategy (Jobs' hope that consumers would buy Macs in order to make their consumer-electronics devices usable), Sony's catastrophic missed opportunity to turn PlayStation into an app and Web-browser platform https://news.ycombinator.com/item?id=22288538 , and the fear other companies felt towards MS even after Apple had been first to a hit smartphone.
It seems pretty clear that in almost all cases, what was most sorely missing was someone with both the power and the taste to make an overall assessment of UX and say "no, this is shit, come back with something acceptable"; maybe to even make a few hirings or acquisitions if necessary to get someone competent on the problem.
By all accounts I've read, that someone was Steve Jobs. After Rubin saw what iPhone was, he apparently completely pivoted Android to go the same way. Before aiming for the full-handset touchscreen paradigm, Android was supposed to be able to handle all types of handsets, including foldables, keyboard slideouts, Blackberry paradigms, etc, all depending on what manufacturers wanted to make. Seeing the iPhone changed the game, so much that Steve Jobs had his "going thermonuclear" rant about Android stealing his stuff. Ironic for a guy who once claimed that Apple was shameless about stealing other people's stuff because great artists steal.
Android could and did handle all kinds of handsets. It wasn't Android that took the industry in the iPhone form factor, it was the OEMs. People tend to think Android had way more power over the mobile industry than it actually did back then. Android got all kinds of dumb stuff put in it because a major carrier or OEM i.e. customer wanted it there.
That also constrained Android's innovation significantly. Some people say Android copied iOS, which is only partly true, and of course iOS has also copied Android over time. But when I was at Google I did encounter quite a few stories of cases where the Android team came up with something really clever and it was shot down by carriers and OEMs who said "we want what the iPhone does". They had no vision at all.
In fairness to carriers though, T-Mobile's QA effort on the G1/G2 were pretty intense. At the time Android's QA was near non-existent and the carrier testing procedure found tons of bugs.