Actually ES9018 isn't operate on that high a voltage. Most DAC can run on slightly higher voltage, but they are also fine when working on 5V and below. ES9018, for what I know, runs on 3.3V and 1.2V (for analog and digital parts of the DAC respectively). Desktop class DAC do however draw more current than mobile DAC, and they also require much better regulated power supply to achieve higher SNR, which of course all add to the size of the circuit - that which comes to the first thing that separates your average DAP and a high-fi DAP: the power supply.
To really achieve a desktop / almost desktop performance on a DAP, you must have a great power supply. It needs to be fairly clean, stable, multi-rail and high voltage. So you need to have switching power supply to pump up the voltage, split it into multi rail (1.2V, 3.3V and / or 5V for the DAC, plus higher dual-rail for DAC output and the amp section), then use regulator to smooth out the ripple and stabilize so it doesn't drop when battery is depleted . Doing so does waste a big part of your power, so you need to use a larger capacity battery to compensate. Now that you have a good performing DAC, you need fairly high voltage so the DAC output or amp section can have a good voltage swing, especially if you are not using rail-to-rail opamp since a lot of the high performance audio opamp don't offer rail-to-rail and rather like to operate in higher voltage. Then you will likely need to put in some buffer after the opamp as those opamp don't usually have high enough current output on their own to drive headphone directly or simply have a high output impedance that need to be isolated. So at the end you might end up having 6~7 chips to reach the level of sound you can be happy with on a high end DAP, all which take out space and power on the player.
So now that you know why high end DAP tends to be big, you might want to ask why can't someone make it smaller? Well, you can, if you can use highly integrated chip and design - that is, if you have that kind of money to use customized chip to your spec. Unfortunately that will be too rich for most DAP maker. Apple can do it but I doubt they are interested at all, given how pricy chip manufacturing can go. So what other way to do it? well, you make compromise, like Sony and iriver do. NW-ZX1 and AK100/120 are taking the easy route by simply using a dumb down amp section that don't go anywhere near the desktop level performance, or using a more mobile friendly DAC that doesn't require as much on power supply. The reasoning is simple: for the majority of user out there, you probably don't need that kind of performance given how easy your typical headphone is to drive. Then again, this does mean that instead of an all-in-one (near) desktop level DAP, it is going to be like a 'half-bake but get most of the job done' kind of DAP that is more suitable for portable user. So it really is a question on how far you want to go with your DAP design. Don't forget, if you need ZX1 or AK100/120 to pump out serious power, you still need an portable amp - then you will end up having a rig just as big as DX100 / HM901, if not bigger. But those are really for the hardcore audiophiles, not the average-consumer-that-have-a-higher-taste-in-music whom Sony / iriver are trying to sell their DAP to.
To cap it off - unless there is something very revolutionary happens on the hardware side, I don't think you will see much change on how high- / near high-end DAP is going to change in the foreseeable future. They either will be bulky and full featured, or they will be slim but missing maybe one or two things.
Edited by ClieOS - 11/30/13 at 6:09pm