The rejection of dataless information in accordance with the General Definition of Information leads to the following modest thesis of ontological neutrality:
There can be no information without data representation.
This thesis can be, and often is, interpreted materialistically, with the equation of representation and physical implementation leading to the following:
There can be no data representation without physical implementation.
which leads to
There can be no information without physical implementation.
This is the view of Landauer. A related discussion of the matter can be found here.
Now, there are few stipulations regarding the nature of data. If semantic information is identified with propositions, which are taken to be immaterial, abstract entities, then perhaps one could possibly say that propositions are made of immaterial, abstract data? Although a datum (lack of uniformity) is instantiated in a physical implementation such as a marked sheet of paper, ultimately lack of uniformity is an abstract phenomenon. For example, the strings ‘110010’ and ‘BBAABA’ would both represent the same data.
This view does seem at odds however with the standard view of data, which sees data as being a physical phenomenon, something that is physically instantiated. If information requires physical implementation, then the strict identification of semantic information with immaterial, abstract propositions needs elaboration. The semantic information that ‘Mars has two moons’ (data say, in the form of a string of English characters) is strictly speaking not just the proposition that Mars has two moons. Rather, it is the meaningful, well-formed data that represents the proposition that Mars has two moons. When two pieces of meaningful, well-formed data represent the same proposition then and only then are they the same piece of semantic information.