(This is another thing I found myself writing on Quora and wanted to keep. The question was "Does Python have any scalability limitations?")
"Scalability" is a term people like to throw around, but the less specific you are as to what you mean by it, the less substantial the answers will be. It is not a simple linear measure on which languages can be given some numerical score.
Languages and their implementations do have certain inherent performance characteristics, but in order to understand their relevance to your needs you have to get specific about your needs.
You will always be able to find stories where people used technology X and found it "didn't scale". Sometimes this is because they didn't know, or chose not to pursue, certain performance optimizations. Sometimes this is because technology X was a poor fit for their problem.
(If you're into programming language esoterica, the technical feature of Python that gets the most attention in performance discussions is the Global Interpreter Lock.)
If it suits your project, use it. There's not some hidden performance ceiling that's going to suddenly appear and crush you. The Python system I work on serves 20+ million pages per day.