A fair point, but wouldn't it only become unrecognizable at levels that mean you're effectively eating pus instead of salmon? My understanding is that the effective innoculation needed to give botulism to a human baby (who has an immune system, just less of one than we do) is <100 spores, which is picograms.

There's just such a gulf between the prices at which this is feasible for food use, and the prices at which existing large bioreactors can culture animal tissue.

If we can't even get plant slop ("algal biodiesel") culture consistent and cheap enough to burn in an engine, or get plant slop ("tilapia feedstock algae") cheap enough to industrialize to outcompete chickens... I don't know that I'm comfortable eating bioreactor meat that can only survive in the FDA danger zone.

Living animals with immune systems are the only types of organisms which can effectively host pathogens such that they can be communicated. Even your example belies the problem: pus is produced by the immune system destroying bacteria, it isn't a bacterial colony itself.

In an a bioreactor where no immune system exists, there can't be a latent infection: there's no immune system! If it can infect and destroy what's growing, then it'll infect and destroy all of it. It isn't going to look like tuna meat after that.