Microsoft reputedly has of a poor record of commercializing their excellent research (they're not alone - e.g. Xerox PARC). It's tricky, which is one reason startups often overthrow the oldguard in the next disruption (as Microsoft itself did).
But your example of Kinect is promising... even if they bought it rather than develop it themselves (idk which), they have commercialized this research.
But your example of Kinect is promising... even if they bought it rather than develop it themselves (idk which), they have commercialized this research.