If the A5 meets it's efficiency specs AND matches up to traditional Innosilicon RELIABILITY, it might be the ONLY X11 miner that will be profitable 6 months from now unless you have VERY VERY cheap electric.
WAY too much hashrate (especially out of Bitmain D3 units) going to be hitting the network by the end of November to leave much room for profitability of less efficient miners.
I'm also not impressed with what I've been seeing so far about D3 reliability - seems to have the SAME issues as the S9 on that score so far.
The A4 met it's FINAL advertised hashrate - it was the early "before the units actually exist" estimates it didn't meet.
On the other hand, Innosilicon really needs to get to work on it's ordering process AND it's customer service - they do a better job than most on supporting their gear, but it's a PAIN dealing with them at all.
I had issues with an A4 preorder from a month or two back - they JUST YESTERDAY finally emailed me with an offer of a fix for the issues, long since I'd SPENT the money on other gear....
The issue is that because of the ridiculously high cost, the A5 will only be profitable after the 6 months. Those first six months you get it are going to be dedicated to recouping the cost. And the further they push back the promised delivery timetable, the longer that breakeven point is drawn out. Regarding the A4, meeting the final advertised hashrate isn't the same as meeting the advertised specs people used to make decisions on whether to plunk down the cash or not. Look at Bitmain's example where their final spec wound up being over what they originally estimated (17 GH vs 15 GH, and less power draw to boot). I'm not pleased with Bitmain for the amount of product they're putting out, but at the very least I can't fault them for delivering ON TIME and over their estimated specifications.