Context
RDF (Resource Description Framework) is the W3C standard for representing knowledge graphs. It is not exotic; it has been an open standard since 1999, and it is the data model that underpins schema.org, dbpedia and most government linked-data publications.
RDF is to knowledge graphs what JSON is to API payloads, the standard wire format.
Explanation
Three things to know about RDF:
It's triples
Every piece of information in RDF is a 'triple': subject, predicate, object. 'Pump-403 is-located-in Building-7.' That is one triple. A knowledge graph is just a very large collection of triples.
Why this is powerful
Triples are uniform. A graph of billions of triples has the same structure as a graph of three. You don't need schema migrations when the data shape changes, you just add new triple types. Federating two graphs is just merging two sets of triples.
Why Weaver chose it
RDF is the only data model that is both formally specified (so it survives decades) and natively semantic (so relationships are first-class). The ecosystem of mature stores, query engines, validation tools and visualisations is twenty years deep.
When we talk about Weaver being built on 'linked data', what we actually mean is: every information container in the system is an RDF resource, every relationship between containers is an RDF triple, and every query is a SPARQL query against that graph.