Where does the energy for magnetizing a metal come from? I saw a young man's science project.  He asked this question in a different way "does magnetizing something reduce the magnetic field potential of the original magnet?"  His answer experimentally was "no".  This begs the question, was he right?
Magnetizing an object clearly requires energy.  It is now placed into an ordered state from which there are many other possibilities.  Therefore, it has moved from a state of relative disorder to a state of order.  The laws of thermodynamics say that this requires energy.
Where does the energy come from?  


*

*Does the magnetic field strength of the original "permanent" magnet decrease?

*Is there "work" performed (e.g. Force x distance) and therefore the energy comes from the work it takes to bring the non-magnetic material close and then tug it away from the "permanent" magnet?

 A: Have a look at this video  .

An iron bar is used to try to pick up some paperclips or thumbtacks. It is not able to do this because it is not magnetized. The rod is placed in a long solenoid and DC power applied. The rod becomes magnetized and is able to pick up some of the paperclips or tacks. The rod is again placed inside the solenoid and 120 VAC applied. This demagnetizes the rod and it will not pickup any tacks. 

In  this video the energy is supplied by the DC current for magnetization, and the AC current for demagnetization . 
If you use another permanent magnet for magnetizing, the energy will be supplied by the motions  of the experimenter. If a magnet and a non magnetized  iron bar are brought into contact for a long time the energy will be supplied from the kinetic energy of the molecules.
It is a re arangement of existing tiny magnetic moments in random directions into an ordered one.
A: Why does the PM not loose his magnetisation? Look at it this way, lets say this is not one PM but billions of little PM-s that are closely together. Each of them generates a field, if you add up those fields they will create a pretty strong field that keeps the all the small PM-s in line. When I now think about it, I conclude that it is a bit like human conformism. The ones on the fringes are less affected by the field so they can create pockets of resistance to the central field.
Now lets say that you change the magnetisation orientation of one little PM, what happens is that the others will force him back! As long as the bulk macroscopic PM isn't overpowered the microscopic PM-s(on the inside) will be remagnetized to the same direction. It can however happen that a large enough part of the macroscopic PM gets demagnetized(boundaries), so that the total magnetisation falls off. This is of course a simplification, but it isn't so far off. 
In short: The force will be mechanical if the other object is not magnetized so strongly that it would overpower the PM.
I must add that the realisation that human society can be pretty well described by a permanant magnet(while writing this) made me feel awkward.
Maybe you should look at the demagnetisation processes.
A: Unmagnetized iron (Fe) rods can become magnetized because of iron's atomic structure
This book does a nice job and even has pictures: (The Foundations of Magnetic Recording by John C. Mallinson, page 4)

If you study electrodynamics then you know that an electric current generates a magnetic field (this is separate from the magnetic dipole moment "spin" of electrons... any moving charge must generate a magnetic field this is simply due to special relativity)... when you pass current through a solenoid a static and straight magnetic field (like N-S magnet) will be generated within the internal of the solenoid.
Now, no one can give you nice formula of the force / counter-force of a pair of electrons (one involved in current and the other stuck in an iron bar). But we do have the very simple magnetic dipole moment potential energy equation: Take the derivative with respect to displacement $(x, y, z)$ and you've got yourself force.
So force is being applied to the electrons in the iron bar. And a counter force is being applied to electrons in the solenoid: electrons in solenoid lose energy, and the electrons in iron bar line up together, and hence the iron bar becomes "magnetized". You should be able to measure a nice voltage drop, until the iron bar becomes fully saturated.
Another way of looking at it is that, you have a time-varying magnetic field being generated in the solenoid, against which the current has to push, hence the voltage drop (in essence you are charging up an inductor when magnetizing an iron bar).
The "lost" energy isn't lost. It is present in the form of a magnetic field. Which if permitted, can release energy in form of current (or a nasty shock as in igniters).
