I'm working on a module called Excel::Reader::XLSX that aims to parse XLSX files without the memory overhead of Spreadsheet::XLSX (which is inherited from the design of Spreadsheet::ParseExcel).

It isn't CPAN quality yet but it reads large XLSX files with a small and constant memory usage (it holds only one row of data in memory at a time). It also runs at a comparable speed. *

* I'm using XML::LibXML::Reader for parsing but I had to profile and optimise the initial code heavily to get to within 5-10% of the regex parsing in Spreadsheet::XLSX.

--
John.