Well, yes. Loading a big file into memory takes memory. Note that the way you're loading in the file,

  @l_Data = <FILE>;
[download]

is going to incur overhead for each line and for the array that holds references to all of the lines. Depending on your line lengths, that overhead could be substantial.

Compare that do

  local $/;
  $data = <FILE>;
[download]

which slurps the entire file into a single string. It'll still take a fair chunk of memory, with (possibly significantly) less overhead.

However, if your process involves morphing the text of the existing file, you might be better holding an array of lines, since the incremental cost of making a copy of a line is considerably less than making a copy of a 25Mb string.

What can you say about your processing needs? Perhaps there's a better way yet.