Correlation doesn't equal causation. I've seen more failures in 70F UPS rooms... Some equipment did have trouble with high temps, but all new equipment can and should be able to take 80F normal operating temperature with limited excursions up to 85-90F. And the more common correlation is claims of higher failure rate for servers at the top of a rack. In most traditional datacenters, you can find hotspots where recirculation from behind results in a continuous 80+F temperature into the server. When we talk about running a datacenter at 80F, you CANNOT have hot spots - you must have solid partions between hot exhaust and cold aisle intake. This is a key issue. Most datacenters with a 70F setpoint actually need it to maintain 80F intake at the ends of rows or tops of racks in areas with messed up airflow management.

"Servers may be able to operate at 90-100, but they simply won't last as long being cooked compared to equipment that lives at cooler temperatures."

Operating 500 hours a year at 90F (the peak of the allowable range) is unlikely to impact longevity. 100F is outside of the allowable range. Your opinion is contradicted by what IBM, Intel, Dell, Sun, and numerous datacenter owners along with the design professionals at ASHRAE have developed over the course of several years of research and many (mostly dull) hours of debate.

There are special cases, tape machines are glaring examples, but operating a datacenter at 80-90F does not have any correlation beyond old wive's tails with increased equipment failure. Indeed, such a 10F difference in actual component temperature (which is what matters) can occur merely between different manufacturer's case layout or the use of meshed back security rack.

"Why they can't further extract useful energy from this hot water I don't know."

I blame that bastard Carnot personally for this... They could get additional work out of that hot water, but it gets prohibitively expensive the lower your delta T between hot and cold gets. I was all stoked about finding some sort of stirling heat engine to run off some datacenter waste heat, until I worked the numbers and found the annual average maximum therorectical efficiency was under 15%.

F*cking entropy.

"I think the concept is interesting, and it makes me wonder if we'll see more datacenters built in areas of the world more conducive to projects like this in the future."

Already happening in a way. Check out EDS's Wynyard facility. They didn't eliminate the chillers entirely last I looked, but in that climate they could have if they trusted the outdoor conditions and local code officials (open cooling towers are subject to abrupt shutdown if there is a Legionella scare anywhere near by in Europe).

Although the lure of unutilized MW is a bigger pull. It's always nice to site a datacenter where the local utility overbuilt and isn't going to ream you when you ask for a MW or twenty.

"Again using rules of thumb, you can assume that 80% of the electrical power delivered to the computers will be dissipated as heat."

? 100% of the electrical power delivered to the computer is dissipated as heat. It's the law. It will be far less than the nameplate power (that electrical uses), and perhaps 80% of what is delivered to the building (after transformer, UPS, and PDUs), but it all ends up as heat (unless you're splitting hairs about the acoustical energy emissions and velocity pressure in the exhaust, which is small and quickly converted to heat).

You do not need a chiller to operate a datacenter in many environments at all. Based on the 2nd edition of ASHRAE's Thermal Guidelines for Data Processing Environments (which was developed with direct input from the major server providers), you can run a datacenter at up to 90F. Seriously, 90F into the rack. When it comes out the back of the rack, you collect the heat exhaust at 100-110F. "Chilled" water at 81F is more than enough to knock that 110F down to 90F- ready to go back into the front of the rack.

The 81F water can be produced directly from open cooling towers (direct evaporation) whenever the wetbulb is lower than 76F (4 degree approach plus a 1F on your flat plate that isolates the datacenter loop from the open tower loop).

You designed an efficient datacenter, but you're five years behind cutting edge (not actually a bad thing for most critical environment clients). The next wave of datacenters will have PUEs of 1.2 or less and redefine the space from a noisy but cool space to hang out to a hot machine room with industrial heat exhaust design.

I actually just finished a chiller less 8MW schematic design and analysis for a bid. It was my second this month (the first was a cake walk - an extreme Twb of 67F, the second was west coast light conditions).

PS: Secondary pumps? Seriously? Unless you have to boost up to 25 psi to feed a Cray or some other HPC I thought everyone who cared had moved onto variable primary-only pumping. (Sorry, feeling a bit snarky after hitting a 40 hour week on Weds...)

Then learn how to capture the waste heat from a datacenter. In the cooling field, there is a wide range of ways to provide very efficient cooling - all the way down to putting 'em in a tent and just letting the wind blow through (a PUE of 1 - more realistically, a PUE of 1.3-1.1 is easily doable if your mechanical is up to par.) BUT, we are just fighting to throw away all that energy. Does Dean Kamen have a sterling engine for us yet that be hooked up to a generator to recover some of that usable heat? I can give him 10MW at a 40 - 60F temperature delta 8760 hours a year if he can just give me the engine...

I find it interesting that you bring up engines. With some clients now proposing things like 120 kW single racks, we honestly are approaching a computer rack putting off the same waste heat as a small automobile. And dealing with the waste heat needs to be treated the same way. The days of throwing CRAC units against the wall blowing into a raised floor are numbered unless power density starts trending dramatically down - you just can't control these types of loads that way.

YES! This is exactly the approach being used currently in the most efficient (short of insane, no-cooling in a tent one-offs) datacenters today. Design the system to provide adequate space control at the typical outside ambient. Direct water cooling isn't even required, it can be done with large coils and evaporation cooling towers to take advantage of the wetbulb depression. As for energy savings, well on a typical 15 MW datacenter you can save about 6 MW for 8000 hours a year... That adds up fast.

A ground loop is not effective for a continuous cooling load - the ground loop is more a season heat storage medium than a heat sink. A cooling tower system is the traditional approach to achieving water at 55F for free cooling. Actually, we're using air to water coils that control datacenter temperature with 65F water, which can be achieved the majority of the time from cooling towers in many climates.