Even so, it's no secret how conflicted they are about "Docs". It's gone through what, 3 major rebrandings in the last five years ("Drive"? "Docs"? "Apps"?), plus it's more of a traditional, old-school enterprise sale, which means salespeople, budgets, big complicated custom integrations, security policy, etc., all things I don't see as Google's core strengths vs. traditional enterprise vendors like Oracle, Microsoft, etc.
Google's big recent push has been getting more companies onto their cloud infrastructure, which when you think about it, it's kind of nuts how they let Amazon eat their lunch in this area. It's GOOGLE: these guys invent filesystems (GFS/CFS), create their own Javascript compiler (V8), hell they've even created two programming languages (Dart, golang) and a database (spanner) -- how Amazon got ahead in cloud hosting really gives me pause.
In any case, "GCE" as they're calling it (Google Compute Engine) might be Google's fourth billion-dollar business. Mobile is a huge thing: there are tons of phones sold every day, developers love it, but it's still too hard.
I imagine the GCE execs faced a build vs. buy decision for "the thing that'll make mobile development easier", push notifications, and a handful of other use cases Firebase has really nailed. So given the choice between (1) building something in-house, which means they're years behind, have to build a brand from scratch, understand the problem space, and recruit the right team vs. (2) write a big check, it's not a very hard choice IMO.
Source: I lived with a firebase employee, I work in a company that GCE's sales team has visited, I spoke at AWS reinvent last year, and I built a venture-funded company on Google Apps.
EDIT: I guess it's not correct to say it's just Adwords and Adsense anymore, but it's still 90%+ advertising, and having some diversity there would go a long way (see child comment)