There is no hold up, they're here now. Problem isn't necessarily the ability to manufacture audiophile-grade USB equipment. Technology is out there.
The problem as I see it is very simple, and two-fold...The mass market is slow to adopt new technology, especially audiophile types. Second, alot of people have no idea how to use a computer well enough to set up a PC-based system. For example, most know little about digital compression, what it means, why it is important. Just those market conditions alone are enough to slow demand and ultimately r&d production. So at this point, there is little reason for the larger manufacturers to invest-but they undoubtedly will eventually. Although audiophile types may be the last to benefit of course.
Some of the stuff out there at the moment qualifies as just good, aforementioned McIntosh etc..All the Mac music server is is an Escient with a Mac faceplace-since Marantz owns both Mac and Escient. Sound was also very much sub-par out of this when I listened. Wireless is another approach, but you have to have big buffers for music that is streamed uncompressed, and there is just something about no wires that screams non-audiophile. Thats just me...
As far as formats etc...Don't waste your time worrying-Apple Lossles. And as far as storage, don't wait for SSD to advance to the point you can put all your 100's of gigs of music on it. HDD is excellent, and much more reliable these days than it once was. SSD is good, and I look forward to it, but I am not waiting around for advances out of it. It is very expensive to produce, when it fails there is NO recovering anything; so HDD will stick around for a while at reasonable prices per Gigabyte compared to SSD per gigabyte.
Like Paradales said, the time is now, and the two choices IMO are either Empirical or Wavelength, everything is a compromise at this point. Choice is limited for the level of gear you own, so you owe it to yourself to give either of those products a hard look before taking the plunge into server based audio.
Difference between Wavelength (WA) and Empirical (EA). EA is largely a modder, although more of his own stuff is being listed on his site..Everything form WA is custom designed from the ground up-no mods. EA tends to have more answers for the sPDIF inclined, whereas WA is largely pure USB Zero DACs, very little sPDIF, no I2S (native interface for alot of DAC chips). As far as sonic differences between the these two, can't speak for EA, and sonics really have to be tested in your own system.
I can say for my system, the Cosecant easily beat out a $12K Nagra player, among other very expensive players. Best digital I have ever heard.