Cost is the main issue (though not exactly in the way you mean), but control is another.
Prosthetic technology is advancing extremely fast. It's one of the fastest moving areas of bodily research right now, for a number of reasons; not the least of which is that prosthetic technology does have many benefits beyond purely helping the disabled - DARPA has devoted large amounts of time and money to advancing prosthetics and cybernetics due to their ability to agument troops. Not in the Crysis supersoldier manner, but in order to allow troops to carry larger amounts of supplies (munitions, sandbags, etc), and to allow them to accomplish tasks that involve moving heavy loads without causing back and joint issues. This is in addition to helping veterans return to civilian life after losing limbs in battle, which is becoming a greater concern with the threat of IEDs and increased usage of irregular warfare.
The reason is hasn't reached Deus Ex levels is due to the fact that we haven't had the technology to accomplish much that's useful for very long; prosthetics have a long list of requirements. They need to have small control mechanisms and servos, to fit inside a (roughly) human-sized limb. They need to be extraordinarily precise, to mimic fine human movements. They need to sense the environment (which is something only just being made possible). They need to be light, in order to be viable for long-term use. They need to look vaguely akin to real life limbs, both for the patient to more easily accept the prosthetic as being "their" limb, and to work around a world built for human-shaped limbs (for example, door handles are shaped to fit the average human hand, and at a comfortable height for the average human size). The "Luke" arm I linked earlier covers a little of this in the voice in the video.
The biggest issue prosthetics have, though, is control. For an arm, each finger has three joints, bar thumbs which have two. There's the wrist, which has two axis of movement. Then the elbow, with another two. Finally, the shoulder, with three. That's
twenty one degrees of movement (and there's probably a couple I've forgotten), and ignores the many more minor movements an arm can do. Then there's the fact that movement in many of these areas, mostly the hands, are not gross movements, but require extreme precision. How does one control such an incredibly complex thing (this is actually an issue with mecha anime as well - they pull off absolutely impossible movements when given a simple 2 joysticks and foot pedals as an interface)? The Luke arm goes into this with one approach: using a large amount of various other muscles and movements to control the various aspects of the arm. This is valid, but has three problems associated with it. One is that this means those limbs and muscles that are used to control the arm are limited in their other tasks when using the arm. Secondly, it means the control systems for the arm are spread all over the body and require either a harness system, or a multitude of devices all wired to each other. Thirdly, it makes control systems very abstract, with little relation to the actual movement being done. This is a hurdle that can be overcome with training by the patient, but means a longer training period, and depending on the muscles used, somewhat slow control. The biggest advances that are happening right now are in the area of nerve control. This uses the nerves and muscles near the amputated limb, but is also extremely experimental.
It seems to be being used for this guy. This seems to be the most promising way forward, as it's noninvasive, small, and minimises the difference for the patient between their natural limb and their new prosthetic.
This is very much focused on "smart" prosthetics, which require computational power to function. "Dumb" prosthetics, such as most leg prosthetics, are at a much later stage of development whilst still having a lot of money poured into their development due to not requiring many of the aspects that smart prosthetics do to be useful. That isn't to say they're any less worthy of research - leg prosthetics still need a lot of development to become more natural, and more advances in them are being made very quickly.
tl;dr the technology is very new, which is why it both costs a lot and why it isn't very useful yet for many patients. Money is making it into their development, though, and many private enterprises are also working on solutions as well as government agencies. Prosthetics promise a brighter future for humanity in many, many areas, and this is being recognised by many.
