They do have a serious advantage, but a the vast majority of benefits of this decade of research is not developed and kept by Google in-house. Key parts of it are either included in non-core components (sensor hardware, commodity electronics) that are used also in other applications (such as computer vision applications in manufacturing industry robotics) and Uber can simply buy that from other companies; or in publicly available academic research and data resources.
In terms of actual competitive advantage in self-driving cars, I'd say that Google has the following things - (a) a vast store of high-precision mapping and laser scan data that nobody else has; (b) a 1 (2?) year gap in pure research between what they have and what has been openly published by someone; and (c) a bunch of engineering expertise and experimentation experience that Uber doesn't have but some others do -e.g. many major car manufacturers. An impressive gap, but one that can be overcome in acouple years if you're serious about it.