Just my take on it but: Because one has to start *somewhere* and get the ball rolling. The older chips were on-hand and they have working boards for them. As a bonus they get to test the DC at full load.
Newer more efficient chips/boards are in the pipeline but not on-hand so you either have a DC sitting there empty/partly used or ya plug in what you have right now and upgrade when the new designs get into full production.
Sure, but doing that needs money and the difficulty is raising like crazy. They will never make ROI on the "miners" so that means that they have money to burn just like I said. I understand that they need to get the ball rolling and that they need to test the DC, but if the 3rd party manufacturers can get miners shipped in March then they can get 16nm miners in the DC too. There is simply no time to make ROI on this 28nm miners.
Please correct me, if I'm wrong.
Let's assume they
1) produce a 28nm miner for $0.06/GH
2) operate it at 0.36 J/GH (conservativ guess)
3) while having electricity costs of $0.05/kWh (or less)
they should make $0.03/GH per month at the current difficulty, which means ROI in about 2 months.
There will be probably first miners based on Bitfury's 16nm ASIC in wild and their DC end of March, but I guess the real volume ramp up will happen in April.
So it should be enough time, to make some profit with the new 28nm hardware.
If you think there will be a 5TH/s miner being sold for $300 within the next week then maybe you can hit ROI with those numbers. Doubt that will ever happen. It needs to be ready to hit the market at half the price of an S7... today.