The bonding is claimed to be proprietary, but it's a reasonable guess that it's a case of polishing the mating surfaces and then using an adhesive of some kind, there are only so many ways to solve this problem. I wonder if it is possible that at the surface exposed to the water that adhesive could progressively allow water into the joint, which eventually forced two halves apart, especially if the low temperatures of the hotel lobby allowed for some kind of expansion-contraction against the 26C tank water.
The problem with this explanation is this might lead to the tank bursting, but I'd expect to see some complete parts of the tank assembly either where the tank was, or at least on the floor adjacent. That would be even more dangerous than the situation we've seen, because a ten-tonne piece of acyrlic crashing through the adjacent hotel block would probably tear a hole in some of the residences. We didn't see that - the pieces of acrylic are, at most, the size of a large sofa in the lobby. Maybe that suggests that the bonding is not simply an adhesive but a plastic welding process, creating effectively a single continuous sheet of acrylic. In which case, the only failure mechanism available is catastrophic.
I'm also surprised we don't see much of the central tube remaining, as I would have thought that would be safe from a failure of the outer walls. Imagine if someone had been in that elevator shaft when the main tank burst. That could have been deadly.