This might be a silly question and possibly nothing to worry about but I figured I'd ask anyway. I recently upgraded my PC (new motherboard, CPU and RAM) while re-using all the other components. I have 2x 780ti running in SLI. I noticed that GPU1 never hits 99% usage during tests or in games. 98% is the highest it'll go and its usually 3-5% lower in usage than GPU2 which has no problem staying at 99% load during tests or in games. Is this anything to be concerned about? Pretty certain I didn't have this "problem" before I upgraded. The differences from old build are:
 
New MB, CPU and RAM. New SLI bridge (just that normal one that came with asus mobo) and I may have reversed the order the GPUs were in during the upgrade (1 is now 2, 2 is now 1). Also OS went from win 7 pro x64 to win 8.1 pro x64. Full specs below.
 
Asus Rampage V Extreme
i7-5930k
NHD-14 cooler
GSkill ddr4 32gb (4x 8gb) 2666 set
2x EVGA GTX 780ti SC with ACX in SLI
Samsung 840 pro 256gb ssd


Running Nvidias latest drivers.
 
Thanks!

The components replaced in the upgrade were Asus Maximus VI Hero, i7-4770k and gskill 16gb 1600 ram. Nothing was overclocked in old config and nothing is overclocked in new one.
 
Playing GTA5 and Witcher 3 at 1440p on an asus pg278q. 

What you're seeing is perfectly normal.
 
Because the Kepler/Maxwell GPU's boost to varying clock speeds, one card (in your case GPU 1), doesn't have to run as hard to keep up with GPU 2.  So when GPU 2 is running at 99%, GPU 1 on has to run at say 94% to keep pace as an example.
 
I see this with my two 970's in every benchmark and game I run.  My GPU 1 has a little higher boost clock than my GPU 2, so I see GPU 2 hitting 99% a lot and GPU 1 anywhere from 1 to 4% percent lower depending on the load.
 
It's possible that someone could "just happen" to get two GPU's that had identical max boost clock speeds.  In that case, you would most likely see both cards running at 99% if the load was demanding enough.
 
That's been what I've seen running SLI setups since the Kepler cards came out.