I can understand why drain voltage reaches 2Vcc, but how bifilar 10T transformer can increase it further by almost factor of 2 is a mystery.

Because the transformer secondary connects to the primary at the MOSFET drain and, because the secondary has an induced voltage across its terminals of almost twice Vcc, the final secondary voltage feeding the coupling capacitor is double what is induced across the secondary.

Consider if the secondary was connected to GND instead of the MOSFET drain; it would only act as a 1:1 device but, because the secondary is connected to the drain you get double the voltage when the secondary dots are shown as per your image.

In other words the AC voltage on the MOSFET drain is bootstrapping the secondary induced voltage.

Incidentally, there is a waveform amplitude error in that diagram: -

Clearly the AC voltage after the C1 capacitor is subject to the same DC bias as that on the gate.

I can understand why drain voltage reaches 2Vcc, but how bifilar 10T transformer can increase it further by almost factor of 2 is a mystery.

Because the transformer secondary connects to the primary at the MOSFET drain and, because the secondary has an induced voltage across its terminals of almost twice Vcc, the final secondary voltage feeding the coupling capacitor is double what is induced across the secondary.

Consider if the secondary was connected to GND instead of the MOSFET drain; it would only act as a 1:1 device but, because the secondary is connected to the drain you get double the voltage when the secondary dots are shown as per your image.

In other words the AC voltage on the MOSFET drain is bootstrapping the secondary induced voltage.

Incidentally, there is a waveform amplitude error in that diagram: -

Clearly the AC voltage after the C1 capacitor is subject to the same DC bias as that on the gate.

Simulation added: -