Startup Idea Case Study: Server Cooling
This afternoon three would-be entrepreneurs from Virginia dropped by our office to discuss their startup idea. They have background in fluid mechanics and have an invention that uses liquid to wisk away heat from servers.
Their key improvement over previous technologies is that the cooling system is powered by the heat of the server itself, requiring no external power.
These three came be our office because Paul Freet told them we’re into high-performance computing (thanks Paul, it was an interesting discussion and inspiration for this post).
Now we don’t do any hardware stuff at AccelerEyes, but I’ve been around enough supercomputing centers to understand more than these three about cooling. Here is my impression of the server cooling space:
- While people gripe a lot about their server cooling bills, no one seems to want to dramatically change the way they do things. Nearly everyone uses server fans to blow air across metal surfaces (i.e. heat sinks) and uses server room fans and air conditioning to transfer the heat outside the room. Methods for outfitting rooms with the electricity and ventilation to work this way are well known.
- Other solutions involve liquids. Few want to put liquids next to their expensive equipment.
- Some innovative solutions involve non-conductive oils. These are cool, but oil is messy and smelly.
- At tradeshows, Dell, HP, IBM, and others show incremental differences on how they cool. But they all sell the same concept: fan/heat sink/AC. At a recent tradeshow, IBM’s innovation was that fans blow from the ground up vertically with heat escaping through the top of the rack, rather than from front to back.
I don’t believe the server cooling space is viable for startups. Maybe there are other spaces that would better benefit from this technology. Maybe these three guys should give up on starting a business and instead search for IP licensing opportunities. Whatever they decide, it’s great they are doing their homework before starting a business.
Sometimes customer development can lead to a dead end. It’s better to recognize dead ends earlier than later.
What examples have you seen of dead end startup ideas?
From the conversation
Comments preserved from the original post.
Mark S.
Thanks again for sharing your time and insight! After only 3 days of customer discovery through the NSF I-Corps program in Atlanta, we already have a much clearer idea of our realistic options and what might lie ahead. The technology we’re developing is not at all tied to the server market, and this process is pushing us to think seriously about other markets.
melonakos
In reply to an earlier comment
Mark, best of luck to you guys! The invention seems awesome and I know you’ll find a great way to commercialize it.