If your strings are stand-alone, then using anchors (\A & \Z (or ^ & $ unless your using the /m option)) will work fine, but if your trying to detect matches for this regex within the context of a larger string, then you can acheive this using a few zero-width assertions.

#! perl -slw
use strict;

while( <DATA> ) {
    chomp;
    print if m[
        \bC                     # Starting with C preceeded by a non-\
+w char
        (?:                     # and either
            800                 # 800
            |                   # or
            (?: 29 | 35 )       # 29 or 35
            (?:                 # and
                (?: 50(?!XL) )  # 50 not followed by XL
                |               # or
                (?: 00XL )      # 00XL
            )
        )\b                     # followed by a word break
    ]x;
}
__DATA__
This  is the spec for the ABC800. This device is a
This  is the spec for the C800. This device is a
This  is the spec for the C2900. This device is a
This  is the spec for the C2950. This device is a
This  is the spec for the C2900XL. This device is a
This  is the spec for the C2950XL. This device is a
This  is the spec for the C3500. This device is a
This  is the spec for the C3550. This device is a
This  is the spec for the C3500XL. This device is a
This  is the spec for the C3500XLM. This device is a
This  is the spec for the C3550XL. This device is a

__OUTPUT__
D:\Perl\test>junk
This  is the spec for the C800. This device is a
This  is the spec for the C2950. This device is a
This  is the spec for the C2900XL. This device is a
This  is the spec for the C3550. This device is a
This  is the spec for the C3500XL. This device is a
[download]