In short: getting throughputs above 300 Mbps with your setup is pretty decent (but not great) result. Sure there are devices which do better (using similar hardware) and there are reports that wifiwave2 driver, which comes with ROSv7 (it's testing software) enables much better performance on certain devices (unfortunately hAP ac² will likely not be supported by that driver).
hAP ac² features 2 spatial streams which, when used with 80 MHz channel, means maximum theoretical physical layer throughput of 866.7 Mbps (
reference). Given that this number doesn't account for duplex (wifi is half-duplex) nor for multiple access guard time, realistic max throughputs are quite a bit lower, so getting a bit more than half of theoretical maximum throughput is a good result. Then there are higher layer overheads (ethernet framing, IP headers, TCP headers), so application throughput gets even slightly lower. Typical test suites measure application throughput. Some add lower layer overhead, but the maths is not exact as actual overhead might be different from what application assumes. The rule of thumb is that realistic throughput is around one half of theoretical peak physical layer throughput.
If there's any interference inside used frequency band (and wifi devices are usually not very good at measuring and showing it), then the physical layer throughput drops ... either because transmitter has to pause for interference to stop or because transmitter has to use more robust (and slower) transmission modes to beat interference. And having both devices 10 feet apart with direct LOS doesn't cancel the interference, it only helps to make interference effects smaller (and it mostly helps with the second cause mentioned, but not so much with the first cause mentioned).
Surely enough
marketing specifications declare this device to be AC1200 device and number implies it should be capable of transfer speeds of 1200Mbps ... but this declaration contains a lot of BS (as do all marketing figures without exceptions). For starters it sums up throughputs of two independent radio interfaces, normal clients can only use one of them. And second, it sums up theoretical maximum physical layer throughputs and then rounds the sum up (actual theoretical maximum physical layer throughputs are 300Mbps for 2.4 GHz part and 866.7Mbps for 5GHz part, so the actual sum is 1166.7Mbps, not 1200Mbps).