As a benchmark: the DJI Phantom weighs roughly 1kg, and its battery can store about 80kJ of electrical energy. I found a NASA study [1] which estimated that the energy released when a li-ion battery combusts is roughly 2x the usable energy capacity. So let's say that a battery fire would release 160kJ of energy.
In contrast, a balloon capable of lifting a 1kg drone would need about 830 liters of hydrogen, with a mass of 75 grams. That amount of hydrogen, if burned in room-temperature air, would release about 9MJ -- roughly 50x as much energy as a battery failure.
(Of course, most of the hydrogen would burn itself up quickly and harmlessly, assuming the balloon isn't flying near any flammable objects.)
I'm not so sure that kind of energy release would be entirely harmless.
You're also not going to get anywhere near the theoretical maximum detonation velocity if it ignites because of a leak.
My point is comparing hydrogen to TNT is very misleading.