> Each zombie-creating fungus species evolved to match a specific insect, so unique strains have little effect on an organism except for the one they evolved to infect. For example, a cordyceps that evolved to infect an ant in Thailand can’t infect a different ant species in Florida.
It is very plausible that the fungus has been around long enough to have evolved with the insects.
The main reason why a cross-species jump is so unlikely for a fungus which induces specific neurological effects is that the mechanism is not intelligent but tuned to the host species’ neurology. The fungus releases specific chemicals at specific times. In the correct host species this results in the desired behavior.
Trying to use the same chemical machinery to control a species which doesn’t share the neurology almost exactly is like trying to run a Windows executable on a Mac.
Importantly, you cannot evolve the machinery bit-by-bit if your life cycle is dependent upon it. Jump to a species with a different neurology would require such a large number of changes simultaneously that it’s just not plausible when anything less would result in a failure to reproduce.