Short story: I'm seeing about 3.5x better wireless network throughput between the new TC and Airport Extreme compared to my 802.11n laptop connecting directly to the TC. Long story: Just picked up the new 802.11ac Time Capsule and Airport Extreme. The goal is to increase my wireless network throughput from upstairs (computer room) to downstairs (entertainment system) so that I can stream high-def movies from a DLNA server. Initial network tests point to big network throughput improvement. My test used my rMBP to perform file copies using different wireless network configurations. My setup is that the new TC is the wireless access point (but not the network router). The TC broadcasts a 2.4 GHz network and 5 GHz network with two different SSIDs. The AP Extreme sits downstairs to act as a wireless/wired bridge. The AP Extreme is configured to "extend" the 5GHz network which is the default configuration made by the AirPort Config Utility. The components in the entertainment system are plugged into the AP Ex via wired ethernet. My network throughput test is using a Retina MBP. I copied a large file using 'scp' from my Linux server to the laptop. Using the built-in 802.11n 5GHz on the rMBP, the copy completed at 16.5 MB/s (~132 mbps). Next, I turned off wireless on the rMBP and used a Thunderbolt/Ethernet adapter to plug the rMBP into the AP Ex. With a wired connection from the laptop with the AP Ex acting as the wireless receiver, I got 50.5 MB/s (~405 mbps) for the copy. Setup 1: Linux -> GB ethernet -> TC -> 802.11n 5GHz -> rMBP = 132 mbps Setup 2: Linux -> GB ethernet -> TC -> 802.11ac(?) 5GHz -> AP Extreme -> GB ethernet -> rMBP = 405 mbps. I am assuming that the TC and AP Extreme can talk 802.11ac at the same time the TC can talk 802.11n to other devices. If that's not true then the increased throughput is the result of better wireless antennas in the base stations compared to the laptop. But I don't think 802.11n can get to 400 mbps so I think the Base Stations must be talking 802.11ac to each other. Also, I am assuming the wireless network is the bottleneck in the copy operation. I believe this to be the case because the my rMBP can write to its flash storage at about 400 MB/s. The file would have been cached on the Linux side, so it should have been in memory on the server. If others get different results over wireless I can try to perform a different network performance test. The results appear to show that wireless network performance is greatly improved with the new generation of Apple Base Stations.