It's up to the specifics of the application, even industrial machinery quite often uses stepper motors. For laser cutting machines up to ~200W for example it's almost ideal cost/benefit wise, there is only the mass of the head and gantry to contend with and steppers are cheap and predictable as long as the loading is predictable. In fact there are many parallels between the Ultimaker's mechanism and the one you would find inside many laser cutting machines - it's just that one is optimised for the delivery of laser light and the other optimised for delivery of plastic and electricity over a cable.
As said there seems to be a huge variation of what servo actually means to people.
'Servo' in general describes a complete control system that operates via feedback. If the stepper motors that we used had a position feedback system, then this would make it a stepper servo and such things do actually exist for various reasons.
The cheapest way of going servo is really the same way that some RepRaps are done with DC motors and encoder strips. I'm not sure about the exact resolution, but something like 0.08mm is probably within reach for those methods.
With the same drive (belt) ratios the RS-555 equivalent DC motor used in most printers wouldn't quite match the holding torque offered by the current NEMA17 - you would have to go to a RS-755 and also put up with occasional 20 ampere spikes in current when your head really does lock up dead or you have the need for speed and want to negotiate a full reversal at full power for whatever reason. Inkjet printers get away with it because they usually decelerate 'off the page' - not really a facility available for our extruders where we don't have the luxury of microsecond flow control.
Even then there's probably going to be issues related to the overall rotation speed of the motors, which determines the intrinsic cooling available to the windings (limiting long term power at low speeds).