That is pretty impressive but it is not the biggest nuke the US ever produced or even detonated, that would be the Castle Bravo test in 1954. Castle Bravo
had a yield of 15mT. The Russians blew the biggest nuke off as a propaganda stunt in 1961 and called it the Tsar Bomba
, its yield was 50mT. 9mT is still pretty impressive though. If I remember right, a B-52 that dropped one of these was not expected to be able to escape or survive the thermal and blast effects of the weapon itself and if it had ever been used they would have been on a suicide mission.
The problem with big bombs is that so much energy is dissipated instead of being delivered to the target because the blast radius is circular and it takes exponential increases in yield to have meaningful increases in blast radius. The emphasis on smaller warheads and, more importantly, more accurate delivery systems is to be able to deliver more blast energy to the designated target thus making the use of the weapons more effective. You run into this same problem with conventional demo to a degree. We say to calculate using the variable P=plenty but it is not always that simple. You cannot just pack more demo and expect it to work, the demo generally must be stacked in the proper way and normally tamped/covered with something to direct the force of the blast. My point is, more smaller warheads with accurate delivery systems are just more efficient than big ones which are just impressive when they explode.