An electric current will produce a magnetic field that can induce a current, opposite the direction of the current that produced it. If This induced current is in the grounded housing of an appliance it will not return to ground on the neutral and can create a small imbalance.
Capacitors and coils are energy storage devices, so when they are first energized the source will see an inrush of current that does not immediately show up on the neutral, creating an initial and usually an intermittent imbalance. As long as these anomalies are small enough to stay below the GFCI trip level of ~3mA, no one notices, but the more devices you put on a single GFCI the larger any imbalance can become on the total circuit.
The inductive coupling to ground in devices should not add up to a milliamp. Probably not even microamps. And this is not an inrush issue either (though transients are known to trip GFCIs). Anyway, transients are not ever additive. Nor is this a power factor issue.
My garage has a 20A circuit dedicated to "tools" with a single GFCI outlet near the panel. From the load side, it leads out over probably 100' (cumulative) of Romex to 11 duplex receptacles in every direction. Most of these have power strips plugged into them permanently, each of which has tools plugged in permanently (chop saw, drill press, scroll saw, band saw, TIG welder water cooler, generator battery maintainers (2), Monitor kerosene heater, TV, at least 6 table lamps, multiple cordless tool battery chargers (Bosch and Milwaukee), two soldering stations and associated smoke evacuators, lights and vacuum for my sandblaster, a cordless phone charging base, and that's got to be less than half the list). Also plugged in permanently are a 50' extension cord reel, a 30' extension cord reel, three drop lights on 20' cord reels, plus a 20' cord reel with the ratchet pawl removed that provides power to a fan and light mounted to the inside of my garage door (I use them when the door is up, and the cord extends as the door goes down). Many of these devices are plugged into IoT relays as well (these have their own internal switching power supplies).
It is hard for me to imagine any scenario more likely to meet your additive threshold, yet with all this ****, I have NEVER had a nuisance GFCI trip (I do test my GFCIs at least annually, and this particular GFCI was installed in 2005). I have tripped it several times due to real ground faults (likely saving me from Electroboom style results at least once), and I regularly trip the circuit breaker with my 12" chop saw (it's rated 18A, but the inrush is severe). I'll also point out that more modern GFCI chipsets have lower nuisance trip rates than my older unit.
OTOH, theater lighting techs will be aware of how GFCIs most certainly can nuisance trip. But in their case, we're talking about a circuit loaded to capacity with power supplies and lighting loads. Outlets and house wiring do not constitute the kind of loads that may add up to trip a GFCI (and if they do, you have a bigger issue).
...On any electrical appliance (except maybe a transformer) the closer you get the rated current and the longer it sees the heavy load the sooner it will fail as compared to other wiring methods that reduce the total load on the same type of device (in this case a GFCI outlet).
Circuit breakers have a "thermal" cutout. A GFCI does not. The time trip curve specified in UL 943 only reaches out to just under 6 seconds:
