Again, Where is the AT&T Fiber?

u_verse_truckAT&T recently announced that they are expanding service of their GigaPower fiber to 35 more communities, bringing the total up to 56 communities. They say that they have plans to pass 14 million homes with fiber by the end of 2019, exceeding the commitment of 12.5 million homes that they promised to the FCC with the purchase of DirecTV.

AT&T says that they were serving 1.6 million homes and businesses with fiber at the end of 2015. The FCC agreement requires them to be offering fiber to 2.6 million customers by the end of this year. AT&T press releases and quotes made to the press claim that the company is already out building fiber like crazy. But are they?

I spent some time on the web looking for evidence that AT&T is building fiber. I started by seeing what I could find about CenturyLink’s fiber build since I know that they are building to pass about 900,000 new homes by the end of this year – about the same goal as AT&T. I skipped over corporate press releases and instead searched for local evidence that CenturyLink is building fiber. And I found plenty of evidence. There are postings by cities warning of coming traffic delays due to construction. There are people posting in local chat groups about CenturyLink fiber. There are newspaper articles taking about the fiber construction.

There was evidence of CenturyLink fiber construction in a lot of markets including: Mesa, Chandler, Phoenix, Scottsdale, Gilbert, Peoria and Anthem in Arizona; Portland in Oregon; Seattle, Tacoma, Vancouver and Spokane in Washington; Salt Lake City in Utah, and St. Paul and St. Louis Park in Minnesota. That’s the kind of aggressive fiber construction needed to pass 900,000 homes in a year.

I found a much smaller list of cities where AT&T seems to be building fiber that includes: Raleigh, Cary and Charlotte in North Carolina; Austin in Texas and Overland Park, Lenexa and Prairie Village in Kansas. These are all markets where Google is also building and where it’s been reported for a few years that AT&T is building fiber. I couldn’t find any evidence for new AT&T fiber construction of any magnitude outside of these Google markets. To give AT&T some benefit of the doubt, perhaps they will be able to meet their 2016 goal to add 1 million passings just in these markets.

I also investigated AT&T’s plans for capital spending. I looked at what AT&T told Wall Street about their capital budget. The company has a $22 billion capital budget for this year. $10 billion of that is aimed at overseas spending including $3 billion to build out from their new acquisition in Mexico. The rest of that spending is aimed expanding its Network on Demand and dedicated Internet of Things (IoT) networks in Europe and to roll out new features for its AT&T NetBond offering in Asia-Pacific, EMEA, and Canada.

Most of the rest of AT&T’s capital budget is aimed at improvements in its wireless networks. The company did tell investors in November 2015 that they planned to spend $2 billion per year for the next three years (2016 – 2018) on wireline networks. Further, AT&T told its investors that it expected overall capital spending to start dropping in future years.

But it’s the wireline capital budget that has me scratching my head. Certainly spending $2 billion in 2016 is enough to add the 1 million new passings they are claiming for this year. But AT&T’s overall goal by 2019 is to go from 1.6 million to 14 million passings. Conservatively that is going to cost at least $12.4 billion over four years just for the fiber. And assuming even a modest take rate for getting 20% of those passings as customers would add at least another $2 billion. Looking back at when Verizon was building FiOS we saw the same sort of big numbers for fiber construction.

It’s just hard to see that AT&T is serious about actually meeting the fiber targets it promised to the FCC. To meet their goals will cost something in the range of $14 billion, and yet they have told Wall Street they will only be spending $2 billion per year on wireline capital. Something isn’t adding up.