I just checked the following address for purposes of an example:
1118 Olive Branch Lane, San Jose, CA 95120
On the AT&T site, that address shows 10mbps and 50mbps available.
On the Sonic site, that address shows 10mbps and 20mbps available.
The same is true of some other nearby addresses that I tested in the area that are within FTTN range.
That is, at least in those instances, the Sonic site does not show that the FTTN speed is available.
This is a location where Sonic can deliver our own Fusion service, POTS and ADSL2+, over copper. And also, AT&T has their U-Verse there, likely using VDSL2 from a nearby cabinet.
What this exposes is how Sonic selects and presents from the potential available services. At a given location, we might have our own fiber, and also access to AT&T fiber, and also Fusion copper - but of course we'll only show the preferred service, Sonic fiber.
In other spot (as in this example), we might have Fusion from the CO, and also access an AT&T IP Broadband offering. Clearly our Fusion single pair beats theirs, but their cabinet-derived product "beats" ours on performance. And if our product was <10Mbps for the X2 offering, we'd flop over to the AT&T product. But when we can deliver 20Mbps, "on net", we consider that superior to 50Mbps over a resold network.
It's left as an exercise for the reader to determine if more megabits off net are better or not, and it's a rather arbitrary best guess business decision our side about how those break-points are selected.
BTW, you'll also see this in some very slow regions - where we _could_ offer an AT&T IP Broadband product at 3Mbps for example, but we just don't consider that a viable product for a new connection today, so we'll simply show no availability at all.