I think they are doing this for a number of reasons. Cost would be one of them. When your company is as technically competent as Google moving your servers to a flavour of Linux and transitioning your workstations to Macs for non-tech staff and Linux for tech staff seems pretty sensible. The money you would save on annual licensing fees would be fairly significant. It is also a symbolic business-driven thing because they are coming to a point where Chrome OS is going to be relatively ready soon. I am not sure why they haven't already transitioned most staff over to some sort of Linux or Chrome OS driven semi-terminal setup running internal Google docs servers for their office apps and then have Macs for those people who need to run say, design programs and so forth.
On the surface the choice to go to Linux boxes or Macs seems counter-intuitive due to their ongoing public fight with Apple in the personal mobile space. But in the business space Apple is not their main competitor, Microsoft still is. MS owns business oem stuff almost wholesale, and that is a freaking huge market. One of the reasons why MS is so ubiquitous is that people work with it every day in the business space. Now that Google have almost all the pieces available right now to run everything in-house they could (and I think, are trying to) turn their own company into an advertisement for their products, to show people that business machines don't need to run Windows and Office in order to be usable. I think that Chrome OS is not only aimed squarely at home users with simple needs, but also mainstream business users. It is a way for people to tap into Google products (internal or external versions) without paying for a Windows licence and without having to deal with the intricacies of Linux.
Now, before people start saying stuff like "oh but you can already do a business machine without Windows! Install Ubuntu with OpenOffice for productivity and Firefox for browsing and Thunderbird with the Lightning calendaring app ... (etc)" just please stop for a moment. Changing people over to a new OS and a new productivity suite and a new browser and getting them away from Outlook for messaging and calendaring all at once is not going to happen without a major effort. At a small office I worked at for a couple years we moved people over to OpenOffice, Firefox & Thunderbird+Lightning while still on Windows and got nothing but bitching about the loss of usability, functionality, and most importantly, familiar a messaging and calendaring system. Perhaps Thunderbird has improved since I last saw it, but back then it was good, but different and finicky about the types of emails and meeting requests it'd receive from Outlook users, even with a number of compatibility add-ons. People don't want to deal with that, they "just want things to work like Outlook does", as they would tell us, constantly.
Outlook on Exchange is an absolute cornerstone of most businesses. Everyone just emails shit to one another. Like, all the fucking time. I have seen dudes sitting next to one another who share a drive send an email to one another instead of saving it and telling the other guy where to find it. Why? It's easier. I know that Thunderbird gives you that same functionality, but it falls over for calendaring, which is something else that the top-level business users depend on heavily. Every workday there are people in my office that are out of meetings for maybe two hours tops, scattered across the day, including lunch. Without a proper, familiar, easy-to-use calendaring system that ties directly into the email & calendaring system on their phones these people would lose productivity, and they are the top-level execs who make decisions about what the company does. If we moved them over to a sub-standard calendaring system they would almost certainly demand Outlook back within a couple of weeks.
Of course Google apps are gonna run up against the same problem, but they have a huge headstart in the area of messaging due to everyone and their mothers having a gmail account and being familiar with how it works. Transitioning people from Outlook to gmail would be a lot better than Outlook to Thunderbird, because people have already used it and know it is something completely different. A lot of the time we would go out to a Thunderbird problem call only to find that the person didn't even realise they weren't using some version of Outlook, and "where is X feature in this version of Outlook?!?!?!" Because gmail is significantly different but still familiar those sorts of problems would not occur as much. The biggest, biggest, hugest problem with moving from Outlook to Gmail is calendaring. Gmail's calendaring still sucks. If they improved it then I'd see this as being a more realistic thing, but until then business users will still love and demand Outlook.