I used to do a lot of firing of clay pieces – sometimes we'd go over 2,000 degrees, which may be a similar temperature range, tho we did it a lot more gradually. Two things would often be destructive:

one was moisture pockets as previously mentioned, altho those usually just showed up as masses that contracted at different rates and eventually cracked and separated from each other.

The things that would really be explosive – usually taking out other people's pieces with them – were certain kinds of little rocks that had snuck through our filtration process, because they were a lot denser than the clay and didn't shrink with the heat as the surrounding clay did, creating a lot of tension that would be released explosively. certain kinds of rock in concrete might have similar properties.

Considering the rapidity of the temp change in a rocket launch, steam is still the more probable of the two theories, but either is plausible.

If I want to speculate – and I do mean speculate – I'm guessing that there was liquid water somewhere it wasn't supposed to be. Subject that to the massive blast of intense heat from the engines, and you could have explosive expansion of steam. (I used to work in a steel factory, and if a blob of molten metal landed in a puddle, it went off like a bomb, shaking the whole factory.)