It wouldn't be a problem until the fan stops, but i makes no sense to keep it in there, and you wouldn't want your PSU to drive away like a submarine

Most if not all server PSU has fan checks and wont work if you simply disconnect the fan, hence a fan simulator or other way to disable the fan checks. We have close relationship with our PSU manufacturer, so of course we wouldn't waste time doing this manually for our DataTank production installations, it is done directly by the manufacturer.

We are not fans of fans, so no PSU or any other fan inside
On a more serious note, depends on PSU's, but via firmware is one way, or hardware modification, but in this case for the purpose of our tests we used a fan simulator.

From the DataTank Mining Prospectus
http://www.datatank-mining.com/files/DataTank_Mining_Ops_Prospectus.pdf

You can technically purchase as little as one unit (0.01kW) of a DataTank if you like, so no its not aimed only for big buyers.

[HAVELOCK] DataTank Mining: 1.2MW 3M Novec Immersion Cooled 2PH Mining Container
https://bitcointalk.org/index.php?topic=655464.0

hehe, that's one way to call it I guess. Its actually sold by weight, but yes if you buy a small quantity, I am sure its pricey.
I Like your drawing btw,

No problem.
The Immersion Blades are immersed in a dielectric fluid, so no damage to electronics.

It dries in matter of seconds and is clean, so yes it comes out dry.

Compared to oil immersion cooling were your hardware would come out oily and has to be cleaned with detergent. Not to mentioned the protective gear you would have to use.

Some ASIC stats after playing around with it for a while, can most likely push it a bit further, enjoy.
Once again, this is in no way optimal conditions for the boards, with the heat dissipation surface down.

http://i59.tinypic.com/2qan7kw.png

I am pretty sure same goes for the BE200 ASIC, but I leave that up to you engineer guys. My comment was of course not meant be negative in any way, just explaining the situation of this tests. I also intentionally kept the ASCI side up to make it look more interesting then a big copper plate. Hopefully soon I can show some boiling action of both sides in a better way.

Quick introduction, since this will be my fist post here. I am a work colleague of antirack, and I am the one that have been doing the setups for this "lab" tests.

I have been running the SP10 for quite a while now, it works great, just not very optimal circumstances for the boards as they had to be placed horizontal and heat dissipation side down. Hence also a bit tricky to get a good recording of the action, if you so will.