Well, there are a few reasons. First, the market for that card is corporations (generally) looking to run the card in the most stable conditions possible. It does have the same chip design as all of the GK110 cards (i.e. Titan, 780, 780 ti, Titan Black, and Titan Z), but it is the highest bin of all. They make sure the card will install and then just work without any hassle, meaning the RMA rate will be very low. Companies are willing to pay a lot for that kind of stability.
Second, the card has 12 GB of ram, so Nvidia will charge a premium for that. Finally, the card's compatibility is much, much higher than any Geforce card. Will the K6000 run games? Yes, and it will do a damn good job. However, it will also do many things that the Geforce cards can't do. Quadro cards support 10-bit displays, so you can reproduce accurate color on very high-end screens, which is important for photo and video studios. Color standards are a big deal in an industry dependent on them. Also, applications will utilize Quadro cards more often. Adobe software, while you can hack the GeForce cards in, will work natively with Quadro. Even if you hack a GeForce, it still can't display 10-bit color. Also, certain CAD applications won't even work with a GeForce hack. Some just plain-old require a professional card.
In short, though, the K6000 costs so much because businesses will pay for them. Companies buy them to be confident that the card will just work, and then keep on working. GeForce cards just can't make that claim, even if they are a little faster.