Beamforming is basically a directional signal transmission/reception technique. Without writing a long reply to explain everything, here's a link
http://www.quantenna.com/beamforming.html
I will quibble with Quantenna's description in two/three areas. One, they don't mention it doesn't really extend range, but it can significantly increase signal strength at even long range, it just won't EXTEND the maximum range really at all.
Two, 12-25dBm increase seems massively optimistic. I have seen a MAXIMUM of about a 8dBm gain in signal strength...which is still a very significant gain in signal strength. Typical is not that good, though it might work better if I had a 1:1 client and my base station was 4:4 instead of 3:3.
Three, the article also doesn't mention that both the client must support beamforming (ALL 802.11ac clients do. Almost no 802.11n clients do, as it was an optional spec and pretty much no one implemented it in 11n). Also, the fewer spatial streams the client has, the higher the possible gain. This is because when beamforming is done, you are combining the timing and frequency of each radio chain transmission to get the signals to form constructive interferance in the direction of the client device to increase the radio signal. The issue though is that some of the signals will suffer deconstructive interfence. So what you might get with a 4:4 router is one spatial stream that sees +8dBm of gain, one with +6dBm of gain, one with -4dBm of gain and one with -7dBm of gain. Or if it changes it around a little, you might see one with +11dBm of gain and the other three with a loss. Just depends on the kind of client it is connecting to.
Anyway, you can only do beamforming if the clients has FEWER spatial streams than the base station. This is because there will be spatial streams that suffer signal loss. So a 3:3 router connecting to a 3:3 client will not perform beamforming, only if it was connecting to a 2:2 or 1:1 client, and the fewer spatial streams the client has and the more the base station has, the higher the possible signal gain.
Again, this only really works at medium and long ranges, but does not effectively extend the range. It can significantly increase performance at long range though.
Example, I have an N600 router and an AC1750 router. Connecting on 2.4GHz to prevent there from being too much bias, the AC1750 router is only a bit faster, about 15% faster same room and only around 2dBm higher signal strength. At medium distance though with a floor, around 30ft and a couple of walls in the way, that N600 router is at -70dBm of signal strength and that AC1750 router is at -60dBm of signal strength. That is -10dBm higher...but maybe only 8 of that is from beamforming as same room, the AC1750 was already 2dBm higher signal strength. At that location in 40MHz mode, the N600 provides 4.5-5MB/sec of Tx ability to my laptop (with a 2:2 802.11ac wifi card in it). The AC1750 on the other hand provides 8.5-9.5MB/sec of Tx ability, a difference of nearly 80% more throughput!
Same room, it was only 15% difference, but at medium range it was an 80% increase!
Move somewhat further away, only 10ft, but with a 4ft masonry chimney in the way and the signal strength plumets to -86dBm with the N600 router and it is -81dBm for the AC1750 router...a difference of 4dBm...and maybe 2dBm of that might be from beamforming. There the N600 can manage 2.2-2.5MB/sec with 40MHz mode and the AC1750 hits only 3MB/sec, a relatively small difference and close to statistically insignificant when looking at same room, but still, around a 20-25% increase in throughput, compared to 15% increase same room...and repeatable for both. That might be more the difference in component quality in the AC1750 router for processing a very weak signal than it is in actually higher signal strength.
Move closer, like down in the basement where the router(s) sits (tests were done with the routers in the exact same place, obviously swapping them out), but with a wall and 15ft or so inbetween and we are back down to only a 20-25% increase in performance with the AC1750 router on 2.4GHz 40MHz over the N600, fairly close to the same room performance difference of the two routers, so maybe a marginal gain from beamforming.
So beamforming CAN help out and help out hugely in some cases. The areas/scenarios where you are going to see a significant increase in performance are limited though. I haven't tested every single square foot of my house, just 5 locations, but compared to the baseline same room performance between the N600 and AC1750 routers, beamforming (or maybe better signal processing? My router does not have the ability to turn beamforming off) seemed to increase performance by 5-10% in 3 of those locations, 15% in one location and 65% in one location.
