Microsoft has a long history of letting its hardware partners take all of the risks. That's a fine strategy if you want to sell a bog standard PC. It is a much less workable strategy if you want to make some sort of unique device. In the past, however, Microsoft has been the only game in town when it came to workable off-the-shelf OS software, especially if you wanted to play well with Windows.
So the OEMs took Microsoft's software and did all of the actual engineering to make it actually work. If the device was a flop they were left holding the bag with the unusable hardware and the bill for the engineering effort. If the device was a success, then it was a given that Microsoft was going to shop your ideas around to your competitors. After all, most of the software that made your device work belonged to Microsoft. Microsoft got paid per device sold, and so they were happy to encourage cut throat competition on the hardware side. This guaranteed that there was a disincentive to actually innovate as companies like Dell, that made their money by cloning other people's ideas and squeezing the supply chain until it bled, dominated. Why innovate if Dell and Microsoft are going to make all of the money borrowing your ideas? Every once in a while someone would come up with a new device based entirely on their own software (Palm, Rim, etc.), but they invariably faced lots of pressure and competition from Microsoft and its OEMs.
The combination of Apple's design prowess and the emergence of Google's android have broken this cycle. Apple has the design genius to create entirely new devices that people want, and the existence of Android means that Apple's competitors have a ready-made OS that doesn't require that they work with Microsoft. Now Microsoft realizes that it needs to get into these new markets, but none of its traditional allies are willing to risk working with Microsoft's software. Heck, HP even has its own software for these devices. Nokia is allied with Microsoft now, and in fact, it has bet the business on Microsoft's software, but they are too busy trying to make a Windows phone to be relied on for a tablet.
So Microsoft gets to take its own risks now. It should be interesting. Microsoft learned a lot from the XBox. It is even possible that they won't lose billions of dollars this time.