With apologies to Mitch Hedberg, if you follow the tech industry long enough, you start to get mad at Microsoft. There's Microsoft's iPhone (Windows Phone), Microsoft's Google (Bing), Microsoft's iPod (Zune). Someone needs to tell Microsoft, "man, just be yourself."
Microsoft seems obsessed with coming up for a "killer" for every hot new trend. If some other company's product gets buzz and "relevance", you can expect Microsoft to come along 1-2 years later and try and compete with that product. By the time Microsoft's product is any good, the buzz moves on to something else. I'm waiting for Windows Live to try and remake itself in Facebook's image.
Microsoft is a mature company. They pay dividends. They're not going to compete in terms of buzz, "relevance", or potential for future growth. They're not in the same league as Apple or Google--but in the league they are in, with companies like maybe Oracle and IBM, they're at the top of the league. So stay in the league that you're winning. Remember, the point of growth and relevance and hotness is to get into the league you've been winning for over a decade.
Yes, Apple is even older than Microsoft and is growing fantastically. That's because Apple had nothing to lose. If you want growth, you have to roll the dice and reinvent the company from the ground up. That's easy if you're Apple and you're at the bottom of the league, struggling to break even. Microsoft is a different story.
Microsoft knows which of its products are worthwhile, which isn't necessarily synonymous with profit. (Visual Studio isn't a big revenue source, but you have to have developer tools for your platform to get developers.) Focus exclusively on those.
Handle extensions in two ways. First, instead of chasing every trend, chase maybe one or two opportunities at a time (company-wide) and focus your best efforts on them. Opportunities aren't trends. Once something is a trend, the opportunity is lost. That's not your league anyway.
Second, don't always chase an opportunity by building something. Like any population, you need fresh infusions of DNA every so often or you will suffer from the effects of inbreeding and die out. If a startup demonstrates a profitable business model that can benefit from Microsoft's scale and stability, buy it. Become the best "pooled-risk company management company" (http://paulgraham.com/prcmc.html). Once you buy the startup, don't destroy or assimilate it. Don't require them to build on your technology stack. Don't move them to Redmond. Carefully control the mixture and interaction to protect the startup's culture and DNA. Over time you can integrate more if it will provide a benefit, but some businesses will have to essentially remain separate companies. Embrace that.