The annoying thing is this is only due to "calibration" which can, with some highly esoteric optimisations, be skipped. Depending on the application , this can realise massive gains in peak power, wakeup latency, and even average power. The whole process is hidden in a binary blob though and espressif will not elaborate on it, so it's very challenging to alter.
Wow, jeez, no wonder the USB won't power it...
That's 600 mA at 3.3V. USB should be fine, but of course you need the capacitance to deal with that. Radio PAs can be power-hungry beasts.