I think they are very different.

Knowledge Graph builds on linked data and semantic web technologies to encode knowledge that is served on an efficient scalable platform.

IBM Watson is a system for ingesting large amounts of text and for then allowing natural language queries on the information in the text.

Both are valuable properties.