Early reports are are not good. For every joule delivered to the chamber, it takes 100 joules of electrical power. Heat to electricity is 50% efficient at best. Reports are that with 2.1mj of input, they generated 2.5 mj of output. Taking inputs and electrical production into account, this means 0.6% is all they are getting out vs. what they put in.
These over-unity reports are meaningless, because every damn one of them only measures Q-plasma, not Q-total.
NIF people like to call this the “target gain”, but someone there has been whispering “net gain” to the journalists, in their ongoing campaign of deliberately deceptive hype. But “target gain” isn’t even (pellet output)/(laser output). The denominator is laser energy deposited in the hohlraum; the rest of it doesn’t count. And this deposited laser energy is estimated based on a model of laser deposition—it’s not measured. (At least, this is the way it was the last time I bothered reading a paper from NIF. I got bored with it a while ago.) The modeling codes are classified; no one without a need to know gets to examine them, and they are not well benchmarked. So the actual target gain is likely < 1 in any case.
NIFs purpose is to validate simulations of thermonuclear weapon detonations. The default is going to be classified, it's making something public that would have to be justified.
These over-unity reports are meaningless, because every damn one of them only measures Q-plasma, not Q-total.