My understanding, dating way back to when Intel first introduced the LGA775 socket, was that the placement of the VRM components is no accident: the down firing stock cooler inherently provides cooling for those components surrounding the CPU/socket itself. To this day Intel has maintained this basic design principal.

I'm wondering what happens when we use "tower" coolers (or liquid setups) which don't provide that same push of air past the VRM components. Is it an overstated requirement, or are we cutting into their lifespan (which I grant is probably longer than the boars practical life)?

In an absolute sense, higher temps will cut down VRM lifetimes as well as add a few watts to your power usage (higher temps = higher leakage current = lower efficiency). However, a couple of things mitigate this a bit. The circuitry is designed to work at high temps, efficiencies have gone up (less waste heat = less self heating), and in general mobos are being designed with better heatsinks/heatpipes.

That said, I prefer some airflow over the VRM heatsinks if possible. Either via case fans (especially for liquid coolers) or a tower fan that drops down below the cooler's fin stack a little bit. Look for mobo's that don't skimp on the VRM cooling.

...and most of all, don't buy VRMs made out of pork products.

Thanks for that.

Anyone have any thoughts on the Noctua NH-L9x65? Its a downfiring model, but I wonder if it would really be any quieter than the stock Intel branded fansink...

Hard to say without knowing the CPU used. But, I would get this model if and only if your case limits you to this size cooler. It's the same price as the NH-L12S, and the latter beats the pants off of it in performance.

I do like the NH-L12S, but was still thinking about those VRMs. Of course the NH-C14S is an option, but at over a full kilo in weight, makes me nervous.

According to TweakTown it is substantially quieter, even at full speed and a better performer as well (20C better at stock clocks, 14C overclocked).