The last test placed the R7800 in my downstairs office and the EX8000 in the upstairs living room. This setup is basically the same used in my original over-the-air mesh test process. Like the mesh process, I used the same kitchen location, one floor up and at the opposite end of my home; my worst-case scenario.
The photo below shows the EX8000's location. (No, Chez SNB isn't overrun with huge spiders. Ms. SmallNetBuilder just likes to decorate for Halloween.)
EX8000 vs. Orbi aesthetics
I should note the EX8000's living room location earned an unprompted query from Ms. SmallNetBuilder, as in "What's the black box in the living room?". When I explained that it did the same thing as the white box (Orbi) and asked if she liked it, the reply was a quick and emphatic "No!". I get that NETGEAR is trying to address a different market with the EX8000 and wants to protect the Orbi brand. But I think they should have gone with the higher WAF design. Just sayin'.
I used a Dell XPS13 running Windows 10 Home 64 bit with a Dell Wireless 1820A Wi-Fi adapter as the test client. This is a 2x2 802.11ac adapter using a Broadcom BCM4350 client device. iperf3 was used to run traffic between a Windows 7 machine running iperf3 server connected to an R7800 LAN port and the XPS13. The EX8000 was set up with different extender SSIDs so I could associate the Dell where I wanted.
Tests were run uplink and downlink on both bands with the R7800 and EX8000 using Channel 40 for (2x2) and Channel 149 (4x4) for backhaul. The key for the plots is:
Baseline - Measurement taken with XPS in same room, ~ 10 feet from R7800
Remote No Ext - Measurement taken in Kitchen, associated with R7800
Remote w/ Ext - Measurement takin in Kitchen, associated with EX8000
The 2.4 GHz downlink plot shows the EX8000 improves a 2.4 GHz client connection in the kitchen from basically unusable to about the same as you'd get with the client parked right next to the R7800. Actually with the 2x2 5 GHz low-band radio (Ch 40)used for backhaul, throughput in the kitchen actually came out better than you'd get next to the router.
EX8000 - OTA - 2.4 GHz downlink
2.4 GHz uplink results show similar throughput improvement vs. direct connection to the R7800 (Remote No Ext).
EX8000 - OTA - 2.4 GHz uplink
The 5 GHz client in the kitchen was unable to even see the R7800's SSID, so there is no Remote No Ext value to compare. Even though downlink throughput via the EX8000 is lower than baseline, connecting the XPS13 to the EX8000 yields about 150 Mbps throughput where there would otherwise be none. There was little difference between using Channel 40 or 149 for backhaul.
EX8000 - OTA - 5 GHz downlink
For 5 GHz uplink, throughput using Channel 40 for backhaul is significantly higher than using Channel 149. In this case, since the 4x4 radio is used for client connection, it appears the extra streams improve the connection for the two-stream client.
EX8000 - OTA - 5 GHz uplink
I'm sorry I haven't anything to compare the EX8000's performance with, since I've given away previous extenders tested and used a new test process. But I think it's clear the EX8000's 5 GHz radios can provide a high bandwidth extended connection over a surprisingly wide signal range. Given this is the same radio used in Orbi, we already knew that. But it's good to see the benefits of Orbi's excellent 4x4 5 GHz radio are now available to everyone.
Yes, $230 is a lot to pay for a Wi-Fi extender and it's almost twice as much as NETGEAR's previous top-of-line EX7000, which I haven't reviewed. But the price might come down over time, especially if Linksys prices its similarly-designed, yet-to-be-released RE9000 more aggressively when it hits stores. Since the Linksys is based on Mediatek, not Qualcomm, silicon, Linksys should have the margin to do it.
In the meantime, if you like the router you have and just need to cover a pesky dead spot, the EX8000 can probably do the job better than anything else you can buy.