Think about it: From the 1940s to the 1980s, mainframes were the only viable computing technology. Terminals could be wired to them, but all software ran on the mainframe. What you saw on your terminal screen was just images, pixels, light.
The mainframe was not a robust technology. If the mainframe went down, all terminals went down, too. If the mainframe failed to store data correctly, all data was corrupted or lost.
The advent of the PC in the late 1970s (it really took off in the mid 1980s) was a far more robust system, since each computer stored its own software and its own data. Microsoft saw the change coming, and correctly understood the dynamic of the PC. That is, you needed an OS that could integrate with the basic software (writing, database, email, presentation, Web) seamlessly. You needed software that was common to the vast majority of users (hence, you needed a company that had a monopoly on software). And the OS should have plug-and-play capability with software produced by other companies. Yes, this meant the OS had major security problems, but the trade-off in productivity was worth it.
Now comes the Cloud. The cloud is essentially a mainframe, albeit one accessed via the Web and not hard-wires running on the ground. The cloud runs its own OS, and runs the software. A user's payment to Google or Adobe or whatever company is out there covers all the software you might need (writing, database, email).
With the Cloud, however, you don't need a huge, fancy OS running on your PC. All you need is a simple OS, with a simple browser. All other software will run on the browser, supplied via the Cloud.
Does this sound like Microsoft? No, not at all.
Essentially, Microsoft's massive, ever-more-complex operating systems are dinosaurs. With only one software program running on your PC (any browser will do), the over-complex OS is dead. There will be no need for MS Word, MS PowerPoint, MS Excel, MS Access, MS IE, MS Outlook. All that software will be hosted by the Cloud. Even computer programming, architecture software, image and graphics, and other speciality software can simply be offered by the Cloud rather than stored on your own computer. (Adobe already no longer offers Photoshop as a stand-alone program. You access it via their cloud.)
But where does Microsoft get most of its money? From the sales of that mega-complex OS. From the sales of those five or six mega-complex software programs.
No wonder, then, that Microsoft is desperately trying to shift away from PCs. It knows that its future is limited. It knows that its OS is dead, and if it wants to make money off it then it had better extend its OS to other types of technology -- like smartphones, TVs, blu-ray players, video games, and the like -- rather than continue to make ever-more-mega-complex versions of the OS. It knows that its five or six software programs are going to be dead soon: Just a single version will be offered, run by and on the Cloud. it won't sell 10 million copies of MS Word any more; it will sell a single version, to Google or to itself (for hosting on the cloud). So where will Microsoft's revenue come from? From the sale of technology itself -- smartphones, tablets, padds, etc. Which is why Microsoft is desperate to enter the equipment field, and get out of the software field. (It's why they bought Nokia.)
Microsoft is already terrified. That's because Apple and Google have a decade's head start on them, without the huge sunk costs in in personnel, infrastructure, and marketing. Microsoft is terrified that, as its revenues shrink significantly, stockholders and others will see it as a "failure" -- and it could enter a death-spiral in which it can no longer pay off its debts or service its sunk costs.
That's called "bankruptcy".
And Microsoft is desperate to find a CEO who can navigate these treacherous waters.