I've been living with and loving a pair of AHB2s for about 5 years. Part of my experience includes comparison with some 'big' amps at Duramax747's listening room. I've never dared push any hi fi system as hard as he does. We were driving the CS5i, The impedance curve differs from the 3.7, but both are very low. The AHB doesn't hold up very well under heavy load. Benchmark showed me graphs supporting amp stability down to 1.5 ohms, below which the signal is interrupted. It is safe. But it isn't as agile and authoritative as they claim.
This next thought is presently for the DIY in some of you. I have tested bi-amping Thiels with very good results. Much of Jim's rejection of bi-amp / bi-wire was defensive. It is all too easy to screw things up. But, conscious players using short speaker runs, identical in every way, can get excellent results. I use the AHB2s and my pair of Classé DR9s in stereo mode which effectively doubles their source impedance behavior vs bridged mode. The XO is split to equalize the speaker load between channels. In the 5, I drive the 3 woofers with one channel and the 3 upper drivers with the other. In the 3.7 I would try splitting the woofer vs the 2 coax drivers. The loads are not equal, but they are complementary; lots of headroom is needed for the big voltage spikes in the midrange.
Of course it's not ready for prime time and YMMV. The point is that the purported gains of bridging become effectively moot for most real world amplifiers driving very low impedance loads. Amps working in their 'normal' stereo mode are much happier.