Amplifiers with improved efficiency and output power are described. In an exemplary design, an apparatus includes an amplifier having at least three transistors and at least two capacitors. The at least three transistors are coupled in a stack and receive and amplify an input signal and provide an output signal. The at least two capacitors include at least one capacitor coupled between the drain and source of an associated transistor for each of at least two transistors in the stack, e.g., at least one capacitor for each transistor in the stack except for the bottommost transistor in the stack. The at least two capacitors recycle energy from gate-to-source parasitic capacitors of the at least two transistors to the output signal, which improves efficiency and output power of the amplifier.