Quite a one-sided view
OpenXML, unlike ODF, is not a fairly developed standard inasmuch as it is far easier for Microsoft to implement than it is for anyone else. Arguing that 3rd parties should implement OpenXML on account of the wide adoption that Microsoft's name brings with it ignores the technical hurdles (and thus cost) associated with its use by third parties.
Why is this the case? Because OpenXML is designed around representing in XML the internal structures of Microsoft Office documents -- not around representing documents themselves the best way possible. As such, it gratuitously fails to take advantage of existing standards which ODF leverages -- and adds a great deal of extra work for 3rd-party implementers who, on account of Microsoft using its own proprietary mechanisms for such things as vector graphics, metadata conveyance and mathematical formula layout, are unable to leverage existing libraries and codebases written for handling the existing standard mechanisms for this. Microsoft, of course, already has a mature codebase for handling its internally-developed formats for vector graphics and formulas -- giving it a massive advantage over everyone else in terms of rendering OpenXML. For an example discussing the value added to ODF via reuse of MathML, see http://www.robweir.com/blog/2006/08/math-you-cant-use.html
While Microsoft has claimed that ODF is unfinished or incapable of representing all aspects of a document format necessary, this claim is not nearly so compelling as would appear at first sight. There are certainly items ODF leaves undefined -- the spreadsheet formula format is one, for instance -- but this is hardly a concern, as this means that MS can simply continue using their existing formula format and be conforming. As for other missing features (such as some list-numbering formats which Word supports), it would be reasonable for Microsoft to define an extension for those features (in its own namespace, such that the elements in the ODF-defined namespaces were fully standards-compliant and the document as a whole was still readable with the exception that these elements would not render as expected). In the alternate, it would have been even *more* reasonable for Microsoft to have participated in the working group which produced ODF in the first place (as they were invited to do) and thus ensured that its output met with their needs.
I'll close by pointing to http://consortiuminfo.org/newsblog/blog.php?ID=1857, where the argument is made considerably better than I can here.