Yes and no. The Python interpreter/bytecode-VM is not as fast as it could be, because only a few people work on it. PyPy is great, but not drop in. Lack of funding is the proximate cause, of course, but that's just masking the ultimate cause of lack of interest. Because people use Python as a glue, just like PHP, or Node. To do the heavy lifting you have many options. (Even in Python, you can just use Cython, use FFI to call out to anything with A C interface, like Rust, or C++, or C. Many Python packages do exactly that, wrap a battle tested old native lib.)

That said the Python VM is a well optimized beast, it's very dynamic, you can monkey patch anything and everything at runtime on-the-fly. This is of course what makes it hard to speed up, too many things to "check for overloadedness" when doing anything. A JIT could help, but that's a lot of work. (Hence why PyPy took a decade, and the whole Python C API compatibility problem is still there - PyPy can only JIT the Py parts.)

Anyway. With that much dynamic stuff and with the VM and the language already well tuned for concurrent memory safety its 'async' story is great and performant. (Like NodeJS'.)

So Python is great for orchestrating whatever you need to handle requests. And if you want to scale up, just spin up more instances. (Just like with Java, but you don't have to think much about the overhead garbage collection, because CPython uses simple reference counting.)

Finally, my concern with Python was the lack of static typing, which made any large project hard to manage. (And resulted in a lot of boilerplate tests and excessive defensive coding.) But mypy is great, and with that it's a joy to work in Python. (Just like TypeScript made the JS world exponentially much more saner.)