If you're going to make a fishbowl out of the stuff, seal any surface penetrations, just like you would if it was a through hole, then, even John should be happy, but it still won't leak if you don't. But, you never have Kerdiboard in a finished assembly where it is sitting in liquid water, there's always something on top of it like tile and thinset, or maybe Kerdiband if there's a seam, or KerdiFix. IOW, there's no water pressure, and the foam (aided by the surface layers) will prevent it from leaking.
FWIW, the melting point of polystyrene is 464-degrees F, which was why in a previous post I thought it should survive up to 212-degrees. It seems that is gets plastic at a much lower temperature (but does not melt), and in combination with the bonding of the reinforcement bonding layer, for safety, the max recommended temp for KerdiBoard is much lower. I'll also point out, I indicated I'd neither seen nor been trained on the stuff when this thread came out, and was basing my comments on the datasheets. One reason why I went to the factory for their training class, so I could better understand the product and its potential uses, which John has not done.
Another FWIW, in a steam room, if you can see a cloud/fog, what you're seeing is not steam! Steam is a colorless, odorless vapor and at standard sea level pressures is at LEAST 212-degrees F, and could be many hundreds of degrees hotter - there really isn't any upper limit to how hot you can make steam until it gets so hot, it breaks the hydrogen-oxygen bond and no longer is a water molecule (and that's REALLY hot)! From a practical viewpoint, there's no need for excessive temperatures in a steam room. Below 212, what you see is not steam, but condensed water droplets, i.e., fog, but it does start out as steam in the generator.