As an engineer, albeit one with little money and even less time these days, I'd say this was a do-able one-off. Yes it has its complexities, but no more than some of the others we've seen (I recall the one that someone designed to cut large blocks of foam for car body prototyping).

A 2.5m x 1.2m (8' x 4') bed isn't impossible, we've seen bigger ones. It doesn't sound like excessive rigidity would be critical here if MDF and foam/composite is all thats being cut. I'd like to see some trials of cutting the composite with a 3 or 4mm single flute cutter in a high-speed spindle to determine whether that would work (maybe WebDude could send someone with a suitable router setup a sample of the material to see how it cuts - offers anyone). I do have some concerns about workpiece holding and also what the sacrificial layer underneath would be as a through cut is needed. It may be that a laser would be a better solution for this material but again some trials would be needed to confirm this (wasn't it Mike at TechServe that had a big-bed laser cutter or did I imagine reading that? If so, maybe he could offer to trial cut a sample to see how effective it is)

Some ingenuity around the spindle/head is needed. I'm sure a dual-head arrangement with a pen and a cutter could be designed and speed wouldn't be too big an issue if there was relatively little text (again some calculations on total stroke length of the writing and moves needed would give some idea of the time taken). A more sophisticated solution could be based on a hacked ink-jet printer.

An interesting challenge and one I'd love to have a go at. I'd have thought it was do-able for less than $25k if 80-90% 'home' built.