I have put huge efforts and endeavors into even improving on 1C. Other than going to the dangerous conductonaut, the better solution, over the tried tested kryosheet, is the Kryonaut Extreme (the pink $105 stuff). for my whole air-cooled system and the one I'm building.
All air cooled.
IMHO
The kryosheet, has limitations and has issues with transfer on several surfaces. It is better than most, but after many hours of testing. The sheet is number 4 from the top, for large die CPU and large die GPU.
I have also repasted and used gelid pads (extreme) on both 4090 FE cards for the non main chip contact surfaces such as ram and voltage regulators. I tried the kryosheet on the gpus and the result was 5C higher than the pink kryonaut extreme.
The sheet has its place, but overall, its a unique use case scenario. I'm not interested in reusability, but i do see the value in the part where it would not need replacing repasting. A sense pf permanence. However, in my testing , the sheet is super fragile and I'm not convinced of its re-usability. Delicate is an understatement.
IMHO of course.
I had not done it yet, but was pondering if a hybrid solution would include a slightly dampened (pick an oil or TIM) so that performance is improved on the sheet at the cost of re-usability. If conductonaut didn't scar the surface of cpus gpus and heatsinks, maybe a conductonaut infusion? Where the 'runniness" of the conductonaut is prevented by the ultr thin sheet?
Just musing.
Anyhow, air cooled test and play systems:
Well, I had forgotten to post the upgrade I did about a year ago, now I'm getting ready for dual 5090. So I figured I'd show the barebones of the system minus the mesh face. Core Specs: 2x 4090 RTX Founders Edition 2x AMD 7773x Epyc (128/256 cores), Gigabyte MZ72-HB0 Dual socket motherboard...
www.techpowerup.com
and record score when air-cooled
GravityMark GPU Benchmark
gravitymark.tellusim.com
test platforms including my new tinkering/toy and some pics of my "old" system: