Your intuition is correct. The thought is that the inbound CO2 acdifies the water and creates a less than an ideal environment for the bacteria. There is also the reality that the CO2 won't dissolve entirely and will create pockets in the filter, making noise and other problems, especially if your pump is in the top of the canister. There are plenty of people who do inject right into the canister and I'm sure have good results. I know I used to, and although the accumulated CO2 was an annoyance, beyond that I can't think of anyone or any "study" that concluded that it was a bad thing. Trouble with my setup is that I run an FX5 which moves about 5 gallons per minute real world, and not a lot of CO2 was getting dissolved before either making a bubble on the underside of the lid, or getting blasted into the tank due to pressure (both actually - the bubble would get big enough and the filter would shotgun it into the tank rather dramatically). That also meant that to get to desired levels of CO2 in the water, I wasted a lot of it. With the Cerges, I halved my bubble count and get better solution.
I have never heard of a warranty being denied for CO2 injection. I suppose it could happen but I would be really interested in hearing the manufacturer explain why
sjb1987: 500 gph "real flow" would probably be way too much. I have a Mag 5 which is in that range and I have to throttle it back with a ball valve on the output side, even with the scrubbies in there. Trouble is that manufacturers LIE about their true flow rates so trying to figure out what will or won't work is daunting. I would go for more than you think you need and then design in a way to restrict flow to a useable rate either with a gate or ball valve of some kind. Also think about removing/cleaning these things a bit ahead of time. Priming them is a pain so putting valves/unions on the piping is going to make R and R a lot easier.