I'm not an expert on this but I believe it lies in entanglement. The particles are distant from one another and we want to measure the quantum state of the whole system, say the total spin of these particles. From the uncertainty principle we can measure a property of one of our electrons/particles and in doing so the number of solutions of the wave function collapses, which is as we would expect. (e.g. we measure the position of the particle we no longer know its energy). The odd thing is in entanglement that the other particles in the total quantum system also "know" what measurement was made on our target electron even though there is no direct link between them. In magnetic systems it's the spin we'll be looking at. This is what Einstein and co-workers called "spooky action at a distance". This is about the limit of my hand-wavy knowledge I'm afraid though!

The diluted magnetic semiconductors are promising materials to be applied in quantum computation mainly because the possibility of controlling the electronic spin, and their easy integration with traditional electronics, based on semiconductor materials. Materials such as GaMnAs, allow to generate and control spin-polarized current due to a hole-mediated sp-d exchange interaction between electrons and the local moments of the open d-shells in the Mn atoms and originates a spin-dependent effective

potential. 

In order to produce a quantum computer, we need to generate and control spin-polarized currents, so the DMS materials, due to its peculiar properties, should be a candidate.

For more info please check APL 104 022105 (2014) and references therein.

Magnetism offers quite a lot of possibilities to generated quantum states necessary for the creation and sustenance of computational operators. Electron Spin is largely a measurable quality. If you measure the minutest variations in the electron spin and a lot another magnetic properties, their ensemble will function as the ecosystem for computational operators with quantum capabilities.