I also ran quick checks on a few other routers. TRENDnet's TEW-818DRU V2 had no spec violation for 5 GHz. But for some reason, I wasn't able to successfully find the 2.4 GHz beacon pulse. I know it was there, however, because the SSID showed up on a wireless client.
I chiseled away the goop securing the connectors on Edimax' BR-6478AC AC1200 router and was able to check the 5 GHz radio, which I found did not exceed FCC transmit power limits. But I must have broken the connection on the 2.4 GHz connector I picked at because I couldn't measure a signal there.
Edimax BR-6478AC AC1200 board
You might ask why no D-Link routers were measured. I had only the DIR-868L, having given the others away. But D-Link clearly doesn't want anyone making any measurements in the field since silicone goop was applied to all antenna connectors.
D-Link DIR-868L inside
EnGenius prides itself on producing high power routers, so I disassembled its ESR1750 to get at its two QCA-based radios. The connections for each radio had only a little goop on them and were easy to get apart so that I could attach the power sensor. I checked only one chain on each radio but found power levels well below the FCC limits. I measured 10 - 12 dBm on Channels 36 and 48, 17 - 18 dBm on Channels 149 and 153 and 16 dBm on Channels 1 and 6.
EnGenius ESR1750 board
This is admittedly not an exhaustive study and I'm not claiming laboratory accuracy for the measurements. But the only two FCC limit violations I found in the ASUS routers didn't require high accuracy; they were pretty obvious.
So while I haven't found rampant boosting of transmit power above FCC limits, it sure looks like NETGEAR has at least two examples for its suit against ASUS.
My thanks again to Rohde & Schwarz for the loan of the NRP-Z11!