Perhaps I am misunderstanding, but that sounds almost like a calculation that applies to a vacuum, not a pressurized atmosphere with plenty of "free" oxygen? Perhaps that's the same for the TNT & battery, so maybe none of these calculations are directly applicable to real-world scenarios?
The point is the kg of hydrocarbon is less able to release its 40MJ all at once because the oxygen which is spread out over 4000 times as much volume needs to get to it. The battery contains both ends of the reaction, but the hydrocarbon only contains one.
TnT or nitroglycerin is much more dangerous in spite of much lower energy density because the energy is all in the TnT.
Of course this isn't the only consideration (the fact that TnT releases nitrogen which suddenly wants to be much bigger is important, for example), but it is a reason to hesitate around such a battery and carefully consider whether you want to throw it around at 100km/h.
I could see the concern, though a car is basically always going to require you to drive around with a potentially lethal amount of energy regardless of the storage medium. Also a battery failure does not mean all energy is released simultaneously. A denser battery could result in a smaller battery that could be better protected from damage.
Depends on the battery a bit. Some NMC chemistries are not far off of it. LFP is far safer. Aqueous sodium ion is nit far from LFP energy density and it'll just boil a bit even if you put one in a blender. From the stats and elements involved, LiS is terrifying, but it could be completely fine to put nails through it.