I will agree that OPENQUERY is quicker on the first read only. If you are performing multiple OPENs, you will see much worse performance, imho.
They are good for reading from non-SQL sources (MDBs, DB2, Excel), but the have to establish a connection and read schema information for every call which, I believe, is not the case with the distributed query construct.
-------
at us, very deafly, a most stares
collosal hoax of clocks and calendars
eecummings