There's certainly rules around how much liquidity a bank needs when lending (otherwise they could lend infinitely), but they don't need liquidity equal to all their loans. That's why it's called fractional reserve.. they only need a fraction of the funds in reserve.
Regulatory liquidity requirements are essentially a function of their deposits not their loans (LCR - though some credit is given to incoming cash flows). But I am not even talking about that. From a very practical point of view, the bank will not allow you to draw on a loan if it doesn't have the cash for it.
Only the central bank can create money out of thin air.
To take a very simplistic example, let's say we start with an empty economy, no money anywhere. The central bank creates 100 out of thin air and buys something from an individual. That individual places the money as deposit with bank A. Now bank A has 100 deposit liabilities, and 100 cash (deposit at central bank).
Now another individual can borrow from bank A and place the money with bank B, so now bank A has a 100 deposit liability and 100 loan asset. Bank B has a 100 deposit liability and 100 cash asset (deposit at central bank). This is the money multipler, m2 = 300, 100 in bank A, 100 in bank B and 100 at central bank, whereas the central bank only has created 100 of m1. However bank A cannot make a new loan out of thin air, it doesn't have cash anymore. The capacity to make loans is with bank B, where the cash is.
It is not the case that banks can make loans with no consideration for their funding and liquidity position, just by creating two accounting entries, only the central bank can do that.
