One example might be how Amazon never got into self-driving cars themselves, but then last year bought Zoox.
Given that, I would've thought that the earlier you buy the thing, the easier it is to integrate it with the rest of your company and the quicker you start gaining the benefits of that. Waiting to buy the winner shouldn't be any cheaper than buying all the competitors in the field at the start, because it's no secret how everyone's doing. So the only reasons not to develop in-house from day 1 are if you're not as good at developing things as a small independent company, or there's some other advantage that you're missing, or the option value from the fact that small companies can go bankrupt on their own (but that's only really applicable if the business is capital-intensive).
Their solution is to buy the companies whose technology they want to leverage for their own, similar to what Apple is doing. This is a good tactic, but then creates another separate issue - how do you integrate all of these companies and their tech and employee's into your massive structure?
This has now created a huge bottleneck in that M&A (mergers and acquisitions) are tasked are buying companies and then having to take several years before they can fully integrate them into our company. It is a painfully slow process, but they still believe its faster to do it this way, then to try and build the tools and applications on their own.
I feel like the jury is out on whether this is a good long term strategy or simply replacing one problem for another.
At some point acquiring other companies will affect the ability to hire and retain developers. It might also be a clue that their ability is diminished, or that internal ideas are not being promoted - whatever process Steve Jobs used to bring the good stuff to light.