Hi Paolo,
The plink2 merger will definitely use more than 1 thread, but it also has to deal with a much more general problem (phase, dosage, multiallelic variants), so I don't expect the speedup relative to plink 1.9 to be *that* large, but we'll see.
Meanwhile, what is your use case? Are you just "concatenating" a bunch of plink filesets, where the samples are the same and only the variant IDs differ (or vice versa, the variant IDs are the same and the sample IDs differ), or do you need to do something more complicated? Concatenation is likely to be implemented in plink 2.0 soon.