Many carriers prefer G.fast in buildings where full fiber is expensive to run. AT&T has been very unclear which of their 12M+ planned homes would get fiber to the basement/G.fast and which would get fiber all the way. They've brought fiber to a million commercial locations and I believe more than 10,000 cell sites. It's relatively inexpensive to extend some of that to residences, well under $1,000/home. (The quotes from Bill Smith are from an excellent Alan Breznick interview. Smith, then CTO of BellSouth, led the recovery after Katrina. He's always worth listening to.)
By late 2014, competitors saw AT&T begin a massive "fiber" build in apartment buildings, complementary to their commercial fiber. AT&T engineers played a major role in defining G.fast. The standard was written to the needs of AT&T and British Telecom.
Everyone in the business was developing products, confident AT&T would be interested. I wrote that many of the "fiber to the premises" lines would be fiber to the basement and G.fast from there. I also wrote that anyone who was sure what AT&T would do was uninformed. I inferred that from hearing different things from various parts of AT&T. They hadn't made up their mind.
The March LR interview confirmed the decision still is open. Smith is hopeful that G.fast, "Turns out to be part of our arsenal." Bill added, "The big opportunity for G.fast is in the MDU space," http://ubm.io/1ot2zuh AT&T is deeply involved with the G.fast community, watching England and all the others starting to deploy. The vendors are constantly visiting with the latest good news.
As CEO Stephenson said a while back, running fiber has become much less expensive. Smith tells LR, "We're way, way beyond meeting that target," referring to a $1,000 install cost. But retrofitting fiber most places will always cost more. "We just can't figure out a better way to dig a trench." Lower fiber costs are important around the world. Free in France reports, "It's quicker, less expensive. We've learnt a lot."
AT&T told Wall Street they would begin cutting capital spending after 2014. They had LTE to a claimed 98% of the country, almost catching up to Verizon. Project VIP, which brought fiber to almost every business, was mostly done. They had virtually stopped U-Verse, deciding to go wireless-only to more than ten million homes.
Keeping capex down implies they will not fiber most of their customers this decade. 12M+ homes is one of the largest fiber builds in the West but is well under half their U-Verse footprint and little more than a quarter of their footprint. They may or may not do more next decade, depending on their decisions on 5G wireless.
Fiber home or fiber basement? Yes.