No FCC that would take another year to approve.
No server powersupplies ? Delta makes 2000+ watt jobbers @220v (old stock are cheap)
Case to small/crowded for good heat dissipation (look to KNC) for open case standard.
Not a lot in the design makes technical sense to me, I make em butt ugly, however Function is over the top. (see antminer S1 for good design)
FCC is fast, like 1 month... Less if you exclude it as a proximity device, or limited operation device. (There is even a free service to apply for you, after they test it.)
But that would only be the board, not the unit. Since the PSU and other components are already FCC approved.
As for heat dissipation... You need only fast air evacuation and a slight turbulence at the cooling-fins with slight positive air-pressure for better thermal transfer. That, or an even evacuation, which requires negative pressure and guided air through shrouds.
The shorter the path, the better. (The optimal design would have been the fins going the other way. Going the length, they pull heat from the front, down to the back, which travels quickly back up to the front, through the heat-sink itself.)
The more fins or surface/area the better, on aluminum's thermal-reflective surface. (Sand-blasting the surface actually helps to increase the surface-area and create some additional turbulence for transferal-time. Rolled micro-slots would be better, like the inside of a water-block.)
With open-air design, most of the air-flow passes right over the unit. However, the thermal reflections have more space to radiate into the highly thermally resistive air. Great if you have a decent reflective design, and not a dissipation or transferal design. This design, however, seems to be a dissipation and transferal design. Indicated by the volume of aluminum used and the parallel fins. For that, you want direct flow in and out, with slight positive pressure in the case. (Not an efficient design, but one that everyone keeps using, for simplicity and cost-savings.)
KNC's secret is the heat-pipes, going to multi-finned radiators that all have independent negative-pressure fans, I believe. That is negated by trapping those individual units within an aluminum box. They are not ones to model as a "good design". That is like 90% of the crappy GPU's that just blow the heat back into the CPU case, with the hope that you have enough case-fans to pull out the power. P.S. Case fans create more heat... From the PSU and from the fan stepper-motor itself. It only pulls-in more dust and that stops the effectiveness of having all those fans in the first-place. Unless you operate in a clean-room with a hepa-filter, down to 0.00001 microns.
Glad you are going to work there, I would have done same or more.