Why are the propagation delays simulated by my Quartus software clearly wrong?

I've designed a circuit of 39 two-input AND gates. The output of the previous feeds into one of the input of the next, and the second input has a logic '1' always at the input. I then use a waveform ...