If you have two GFCIs connected in series, which would trip first or both?
All of them
And good luck resetting them.
I mean, it doesn't do any harm, (except in the sense that security at the expense of usability is at the expense of security, i.e. someone tears it out because its annoyances drive them nuts). It just doesn't provide any benefit, and as such, is just a silly thing to do.
Every device is seeing the ground fault, so every device would trip. One might imagine a race condition where one device sees the ground fault first and physically severs power before the others are able to detect it. However there is a mechanical delay between detection and the hammer physically tripping, and that delay is considerable in the time scale of detection. As such, all detect the ground fault before any can actually interrupt. The physical energy to trip is pre-loaded in the mechanism (loaded by you pressing the "reset" button), so it doesn't need power to trip.
If those have been coordinated:
time delayed (the first trips an order of magnitude later than other)
current selective (the first trips for current an order of magnitude bigger then second)
Say the first trips after 1s and the second trips after 50ms or the first is 300mA and the second is standard 30mA.
Then you'll have the second to trip for "little" faults and the first tripping for bigger ones or for ones occurring before the second RCD.
If you put same selectivity/delay devices on a line you'll have a total random behavior so just don't do that.
Unknown - they both trip as fast as they can
The problem is that even little faults can be fatal. So at the consumer (residential) level, I don't think you will find a large difference in trip time. Without a large difference in trip time, mechanical factors (how long does it take to actually "trip") can be longer than the measurement/analysis (which is itself very simple) time. Industrial equipment may have different categories of protection. There might be a difference in some industrial-grade protection.
The key is that GFCI is fundamentally very different from traditional over-current protection. Over-current protection is based largely on heat that could lead to fire. A HUGE amount of heat in a very short period of time can be very dangerous - think "spark" - and therefore triggers a fast trip. A small amount of heat - what you get from putting 30A on a 20A 12 AWG wire - will really not lead to dangerous (fire-producing) levels of heat, so the breaker can be designed to trip much slower. There are variants, and protection is sometimes also to protect delicate electronics or detect and top overloaded motors, etc. so there are different trip curves. But the basic concept is that with over-current protection, a little bit of over-current for a little while is OK (often expected - e.g., large motor startup), but with GFCI a little bit (beyond the safe exposure level) of ground fault is dangerous even over a very short period of time.