The same argument applies to any company with large amounts of cash in the bank: Google, Apple, Oracle. If we're not asking why Apple doesn't do that, then we shouldn't be asking why Microsoft doesn't do that.

One obvious problem is management. Microsoft (and pretty much any other successful company) gets majority of its profits from a very small number of highly profitable products. It's hard to manage several businesses. Imagine how hard it would be to manage 8500 independent business. Or integrate them meaningfully into Microsoft (at that scale, they would have to integrate a whole startup as quickly as they integrate a single engineer today, i.e. several per day).

They would also need hundreds of people involved in picking the startups and making the deals.

For Microsoft, a company/product that does 10 million/year, even if it was acquired for 1 million, will not make a difference. They need billion dolar products.

So the question is: if they invested in 8500 startups, would Skype be one of them? If not, they still would have to buy Skype.