suppose the device is even 1 Megawatt (which is plausible). This is not its continuous power output. I imagine it operating at maximum for a few seconds(1-2) while engaging a high-flying target. Even if it operates for 1 hour, the total energy requirement would be 1 MWh. Using current battery technology, the battery would weigh a maximum of 5-6 tonnes. However, I believe it wouldn't even need this much energy. Even if it engages 300 subsonic cruise missiles in a saturation strike scenario, it would only need to operate at maximum power for about 10 minutes, consuming only one-sixth of the battery power.
I assume most of the bulk goes into cooling, as the power level in the fibers must be very high. On the other hand, a commercial Raycus fiber laser of 220 KW (a more common high-power industrial model), a continuous-wave laser with that much power coming from a fiber only a few hundred microns thick. Therefore, this military laser should be multiple times more powerful than commercially available ones. I think around 1 MW is feasible, and the aperture might have been increased to better deal with atmospheric attenuation and beam spreading.
Isn't it possible for China to combine 4-5 Raycus-type fiber lasers into one unit and add some adaptive optics and other components?