I wonder if you would do better with two hashes rather than two arrays. You could join your array with newlines and open a filehandle against a reference to the resultant scalar. You could then read it in paragraph mode one record at a time and use splits in maps to populate the hashes.

$ perl -Mstrict -Mwarnings -MData::Dumper -E '
my @array = (
   q{CPU Temp = 30},
   q{GFX Temp = 45},
   q{RAM Temp = 40},
   q{},
   q{CPU Status = OK},
   q{GFX Status = OK},
   q{RAM Status = OK},
   );
open my $inFH, q{<}, \ do { join qq{\n}, @array } or die $!;

my( %temps, %stats );
{
    local $/ = q{};
    %temps =
       map { split m{\s+=\s+} }
       map { split m{\n} }
       scalar <$inFH>;
    %stats =
       map { split m{\s+=\s+} }
       map { split m{\n} }
       scalar <$inFH>;
}

print Data::Dumper
   ->new( [ \ %temps, \ %stats ], [ qw{ *temps *stats } ] )
   ->Sortkeys( 1 )
   ->Dumpxs();'
%temps = (
           'CPU Temp' => '30',
           'GFX Temp' => '45',
           'RAM Temp' => '40'
         );
%stats = (
           'CPU Status' => 'OK',
           'GFX Status' => 'OK',
           'RAM Status' => 'OK'
         );
$
[download]

I hope this is helpful.