Let me be direct with you: if someone tells you there's a single 'best' data center cooling solution, they're either selling something or they haven't been in the trenches. I say this as someone who's coordinated rush equipment orders for dozens of facilities, including a same-day turnaround for a financial services client whose CRAC unit failed at 2 PM on a Friday. The normal lead time was 8 weeks. We pulled it off, but not by following generic advice.
Your cooling strategy depends on where you are right now. So let's break it down into the three most common situations I see — and what actually works for each.
If you have the luxury of planning from scratch, this is where you can optimize for the long game. The big question here isn't just 'which technology is best' — it's about total cost of ownership over 10+ years.
Based on what I've seen across dozens of projects, indirect evaporative cooling (the kind Munters specializes in) is often the smartest play, especially in climates where you can leverage outdoor air without introducing humidity issues. What most people don't realize is that standard chillers and CRAC units are incredibly energy-intensive over their lifecycle. The operating cost difference isn't small — we're talking 30-50% lower energy bills in the right conditions.
Here's something vendors won't tell you: the 'standard' cooling solution often comes with hidden costs for humidity control. A dehumidifier running alongside your AC isn't just a second expense — it's a sign the primary system can't handle the full load. A single, integrated system like the Munters Oasis series avoids that entirely.
Don't just spec the cheapest upfront option. I've seen companies save $20,000 on installation only to pay $15,000 a year extra in electricity and another $10,000 in maintenance. Over 5 years, that's a $95,000 mistake.
This is the most common call I get: 'Our current system works, but the power bills are killing us, and we're worried about future capacity.'
You don't always need to rip and replace. In many cases, you can add indirect evaporative cooling as a supplementary system. This is basically adding a free cooling module that handles the majority of load during cooler months, letting your existing chillers idle. I have mixed feelings about hybrid approaches — on one hand, they're not as elegant as a single integrated system. On the other, the ROI often pays back in 18-24 months, which makes them a very practical choice.
The key metric here is hours of free cooling per year. If your climate gives you more than 3,000 hours where outdoor air can handle cooling with indirect evaporation, it's almost always worth it. We've seen facilities in Seattle and Northern California hit 5,000+ hours.
The gotcha with supplementary systems is integration. If the controls don't talk to each other properly, you get suboptimal performance. Make sure whoever installs it specs the right sensors and controllers. Trust me on this one — I've seen a retrofit that 'saved' 20% on paper but delivered only 8% because the control logic was wrong (ugh).
This is where the 'time certainty premium' comes in. If your cooling system has failed or you need capacity now, the right choice isn't necessarily the cheapest or even the most efficient — it's the one that will arrive and work within your deadline.
In March 2024, I had a client call with a data center overheating because their chiller went down. Ambient temps were hitting 95°F. Normal lead for a new CRAC unit: 6-8 weeks. They didn't have that. We sourced a Munters indirect evaporative cooler — the kind typically used for smaller server rooms — and had it delivered within 72 hours. It cost about $4,000 more than standard delivery (and we paid $800 in rush fees), but the alternative was a partial shutdown that would have cost $30,000 in lost revenue per day. An easy decision from my side.
In emergency situations, you want reliability over everything. I've tested six different expedited sourcing options over the years, and honestly, the ones that work best are the ones with dedicated stock or inventory buffers — which is why Munters was able to pull it off. They had the unit on hand and a team that understood 'urgent' means urgent.
They go for the cheapest emergency option. Emergency is the worst time to bargain hunt. You end up with a system that's undersized, or arrives damaged, or doesn't integrate with your existing setup. I've paid for that lesson myself.
Here's a simple framework:
If you're somewhere in between — say, you have 2-3 months lead time — you can still often get a turnkey solution from manufacturers with efficient production lines, but you lose the ability to do extensive customization. It's a trade-off.
Whatever your situation, the worst move is to do nothing. Cooling isn't optional for a data center — and neither is planning. Get ahead of it now, and you'll save yourself a frantic 3 AM call later. (Take it from someone who's answered plenty of those.)