PCR on it's own as a technique is semi- quantitative procedure, because it only amplifies exponentially nucleic acid molecule through a repeated cycle. However, it could be made a quantitative technique by eliminating or minimizing interference to the actual doubling or the amount of target amplicom from each cycle.
Before qPCR machines & reagents were common, folks would set up PCR reactions where they would remove samples after 10, 15, 20, etc. cycles and compare the amplification by running out the products on an agarose gel and looking at the relative band intensity.
You'd never get that type of data accepted for publication now. There was no validation of the efficiency of the primers, it was easy to oversimplify and "saturate" the gel, etc.