Google started work on ChromeOS before Android became successful in the market and long before it arrived on tablets. Additionally, I don't think Google really wants everybody on an OS that runs native apps. Their speciality is the web. Their money comes from the web and the more time people spend searching and browsing the more money they make. ChromeOS is probably what they wish everybody were using, but they aren't going to ignore Android's success.
They'll keep building ChromeOS to hedge their bets. They can't risk getting shoved out of the OS game in any way.
I would have agreed with you if this was the Google of 4 years ago, but I think the faction of the company that thinks web-first is either gone or silenced. Android has the worst support for web apps as first class apps of any of the major mobile OSes.
I had a chance to speak with a Google engineer about this not long ago. They said that Chrome was meant to be cloud-centric, whereas Android focuses more on Apps. (A bit of an oversimplification, but generally accurate.)
My take: ChromeOS may actually be better for Google's interests in the long term. However Android's success has resulting in significantly more momentum currently. The release of Chrome on Android might be an effort to move Android more in line with Google's larger goals.
ChromeOS/Chrome Web Store seems to be about countering the the trend towards "appification" (the consumption of the web through native clients) by turning the web itself into an app store and giving websites native abilities.
Android is Google's take on the appification model. While there's overlap in that Android has a Chrome browser, I think the distinction is that Android's interaction model means that native apps will be primary method for navigation for the forseeable future since the native Web app model doesn't translate cleanly on a touchscreen device.
The Web is essentially Google's Desktop OS, so it has a vested interest in keeping it alive. ChromeOS is a way of accelerating that process, since the closer it gets to becoming viable alternative to an OS running native apps, the closer the web gets by extension. Creating a standalone OS allows them to frame the problem better than just Chrome browser, since a host OS always has a slight masking effect, even if the observer is aware of it.
ChromeOS is based off Linux. Android and Linux are going to merge. Chrome is coming to Android. It's not hard to see imagine some interesting collision scenarios and user experiences.
Once Chrome becomes stable on Android and it replaces the stock browser, I don't know why you'd still want ChromeOS on a separate device, or say even dual booting with Android. Would there really that huge of a security advantage to do that, instead of simply running Chrome from Android?
Because besides the security issue, I don't see any advantage at all for ChromeOS in that scenario. And even so, the convenience of having Chrome in Android most likely trumps whatever security advantage ChromeOS has over Chrome on Android.
